Information Integration Solutions Center of Excellence

Parallel Framework Standard Practices
Investigate, Design, Develop: Data Flow Job Development
Prepared by IBM Information Integration Solutions Center of Excellence July 17, 2006

CONFIDENTIAL, PROPRIETARY, AND TRADE SECRET NATURE OF ATTACHED DOCUMENTS
This document is Confidential, Proprietary and Trade Secret Information (“Confidential Information”) of IBM, Inc. and is provided solely for the purpose of evaluating IBM products with the understanding that such Confidential Information will be disclosed only to those who have a “need to know.” The attached documents constitute Confidential Information as they include information relating to the business and/or products of IBM (including, without limitation, trade secrets, technical, business, and financial information) and are trade secret under the laws of the State of Massachusetts and the United States. Copyrights © 2006 IBM Information Integration Solutions All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM. While every precaution has been taken in the preparation of this document to reflect current information, IBM assumes no responsibility for errors or omissions or for damages resulting from the use of information contained herein.

Information Integration Solutions Center of Excellence

Document Goals
Intended Use This document presents a set of standard practices, methodologies, and examples for IBM WebSphere® DataStage Enterprise Edition™ (“DS/EE”) on UNIX, Windows, and USS. Except where noted, this document is intended to supplement, not replace the installation documentation. The primary audience for this document is DataStage developers who have been trained in Enterprise Edition. Information in certain sections may also be relevant for Technical Architects, System Administrators, and Developers This document is intended for the following product releases: - WebSphere DataStage Enterprise Edition 7.5.1 (UNIX, USS) - WebSphere DataStage Enterprise Edition 7.5x2 (Windows)

Target Audience Product Version

Document Revision History
Date
April 16, 2004 June 30, 2005 December 9, 2005 January 31, 2006 February 17, 2006 March 10, 2006 March 31, 2006

Rev.
1.0 2.0 3.0 3.1 4.0 4.1 4.2

Description
Initial Services release First version based on separation of EE BP into four separate documents, merged new material on Remote DB2, configuring DS for multiple users. Significant updates, additional material Updates based on review feedback. Added patch install checklist item (7.10) and Windows 7.5x2 patch list. Significant updates, new material on ETL overview, data types, naming standards, USS, design standards, database stage usage, database data type mappings, updated styles and use of cross-references. Corrected missing Figure 9. Added new material on establishing job boundaries, balancing job resource requirements / startup time with required data volume and processing windows, and minimizing number of runtime processes. Moved Baselining Performance discussion to Performance Tuniing BP. Expanded performance tuning section. Removed Architecture Overview (now a separate document). Expanded file stage recommendations. Updated directory naming standards for consistency with DS/EE Automation Standards and Toolkit. Segmented content into “Red Book” and “Standards”. Clarified terminology (“Best Practices”). Incorporated additional field feedback.

May 08, 2006 July 17, 2006

4.3 5.0

Document Conventions
This document uses the following conventions: Convention Usage Bold In syntax, bold indicates commands, function names, keywords, and options that must be input exactly as shown. In text, bold indicates keys to press, function names, and menu selections. Italic In syntax, italic indicates information that you supply. In text, italic also indicates UNIX commands and options, file names, and pathnames. Plain In text, plain indicates Windows NT commands and options, file names, and pathnames. Bold Italic Indicates: important information.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 2 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence Lucida Console Lucida Bold

Lucida Console text indicates examples of source code and system output. In examples, Lucida Console bold indicates characters that the user types or keys the user presses (for example, <Return>). In examples, Lucida Blue will be used to illustrate operating system command line prompt. A right arrow between menu commands indicates you should choose each command in sequence. For example, “Choose File Exit” means you should choose File from the menu bar, and then choose Exit from the File pull-down menu. The continuation character  is used in source code examples to indicate a line that is too long to fit on the page, but must be entered as a single line on screen.

Lucida Blue

This line  continues

The following are also used: • Syntax definitions and examples are indented for ease in reading. • All punctuation marks included in the syntax—for example, commas, parentheses, or quotation marks—are required unless otherwise indicated. • Syntax lines that do not fit on one line in this manual are continued on subsequent lines. The continuation lines are indented. When entering syntax, type the entire syntax entry, including the continuation lines, on the same input line. • Text enclosed in parenthesis and underlined (like this) following the first use of proper terms will be used instead of the proper term. Interaction with our example system will usually include the system prompt (in blue) and the command, most often on 2 or more lines. If appropriate, the system prompt will include the user name and directory for context. For example:
%etl_node%:dsadm /usr/dsadm/Ascential/DataStage > /bin/tar –cvf /dev/rmt0 /usr/dsadm/Ascential/DataStage/Projects

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

3 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

..........................................................................................................29 2.......................................6 PARALLEL SHARED CONTAINERS.............................................................................................................................................................................................................................................6 1.....................................................................................................6 COLLECTING METHODOLOGY.........................................................................................................................................................................................................5 COLLECTOR TYPES.............................7 1........3 JOB DESIGN TEMPLATES.............................................................................................2 NAMING CONVENTIONS........................................77 6................................................................77 6.........54 4...............................82 7.............................................................................................................................................................................................................................76 6................81 7.........................................................................................................................................1 WHICH FILE STAGE TO USE..3 DOCUMENTATION AND ANNOTATION........................................... transcribed............................................58 5 PARTITIONING AND COLLECTING....................................................................................51 4 DATASTAGE DATA TYPES...........................................2 JOB TYPES..2 ESTABLISHING JOB BOUNDARIES............................................................................................................................. All rights reserved......................................................................................................74 6............................................................ ....75 6....................................73 6 SORTING.............................................................................................................................................................................................................................................................................59 5.............................................................................................................2 DATA SET USAGE.....85 8 TRANSFORMATION LANGUAGES...............................................................4 PARTITIONING EXAMPLES...............................................................................67 5..............3 PARTITION METHODOLOGY................................................................................................. stored in a retrieval system.........................................................................8 COMPONENT USAGE.................................................................5 UNDERSTANDING A JOB’S ENVIRONMENT....... or translated into any language in any form by any means without the written permission of IBM........................................72 5..............................................................................................................................................................................2 COMPLETE (TOTAL) SORT.....................................................................7 ERROR AND REJECT RECORD HANDLING...........................................................................................................................56 4.................................................................39 3.............................................................3 LINK SORT AND SORT STAGE...............................................................................................................................................................3 SEQUENTIAL FILE STAGES (IMPORT AND EXPORT).......................................................................8 2 STANDARDS...............Information Integration Solutions Center of Excellence Table of Contents 1 DATA INTEGRATION OVERVIEW................................68 5.....................................................................................................1 JOB SEQUENCES............42 3......................................................................................................................................6 AUTOMATICALLY-INSERTED SORTS.............................................................................................................................................1 PARTITION TYPES....................................................................................35 3 DEVELOPMENT GUIDELINES...........................................................................................................4 DEFAULT JOB DESIGN.......................................1 PARTITION AND SORT KEYS..81 7....43 3...................................................................................................................................................................................................................................................................................................................................................................................................31 2.......................87 Parallel Framework Red Book: Data Flow Job Design July 17....79 7 FILE STAGE USAGE............1 MODULAR DEVELOPMENT .4 STABLE SORT...........13 2....................8 TUNING SORT............................4 COMPLEX FLAT FILE STAGE...................................................................................3 RUNTIME COLUMN PROPAGATION..........................................39 3...........................................................................................................18 2.74 6................................................. 2006 4 of 179 © 2006 IBM Information Integration Solutions...............7 SORT METHODOLOGY...........................................2 MONITORING PARTITIONS..........................................5 JOB PARAMETERS.......................................................................................................................................................43 3...................................13 2......................................................................................................59 5.........1 DIRECTORY STRUCTURES......................................................................................................................... No part of this publication may be reproduced.............................................................................................5 SUB-SORTS.............................................................................................................................................................................79 6.................................................................................70 5....................................................................................................... transmitted..................................39 3....81 7...........................................41 3.................................................................................................78 6.......40 3.....................................................................................................................................................4 WORKING WITH SOURCE CODE CONTROL SYSTEMS.........................................................................................................2 NULL HANDLING.............................

..................................................95 10 DATABASE STAGE GUIDELINES..............................................2 DEBUGGING ENVIRONMENT VARIABLES.......................................................... or translated into any language in any form by any means without the written permission of IBM.................................94 9....................... stored in a retrieval system.......................................146 APPENDIX C: UNDERSTANDING THE PARALLEL JOB SCORE..............3 MINIMIZING RUNTIME PROCESSES AND RESOURCE REQUIREMENTS.................................................................103 10................ MERGE........................................................ transcribed.................................................160 Parallel Framework Red Book: Data Flow Job Design July 17....87 8..................................................................2 MODIFY STAGE..94 9.......................................120 11 TROUBLESHOOTING AND MONITORING.....................................................2 UNDERSTANDING OPERATOR COMBINATION..............................154 APPENDIX F: SORTING AND HASHING ADVANCED EXAMPLE................96 10.........................................2 CAPTURING UNMATCHED RECORDS FROM A JOIN....................4 UNDERSTANDING BUFFERING..........................................134 APPENDIX A: STANDARD PRACTICES SUMMARY.4 ODBC ENTERPRISE GUIDELINES.................................................................... 2006 5 of 179 © 2006 IBM Information Integration Solutions.............................................................................................................91 9 COMBINING DATA...................113 10.....................................................3 THE AGGREGATOR STAGE.................................. .........................4 VIEWING THE GENERATED OSH...........................................................................3 INFORMIX DATABASE GUIDELINES................96 10....................................................................................................................................................................................Information Integration Solutions Center of Excellence 8...140 APPENDIX B: DATASTAGE NAMING REFERENCE.............. transmitted..................................117 10............................................................. JOIN VS....114 10..........................................131 12............6 SYBASE ENTERPRISE GUIDELINES...........................................................................................................................124 11...................124 11.............................................................................................................................................................................................................7 TERADATA DATABASE GUIDELINES...........................................................126 11..........................................................................................................................................................................1 DATABASE DEVELOPMENT OVERVIEW............................................ No part of this publication may be reproduced....................................129 12....................................................153 APPENDIX E: ENVIRONMENT VARIABLE REFERENCE.............................................................3 HOW TO ISOLATE AND DEBUG A PARALLEL JOB......................................................................................................................................................1 HOW TO DESIGN A JOB FOR OPTIMAL PERFORMANCE.............................. All rights reserved.................................................................................................................................................................................................................................................................94 9............................................................5 ORACLE DATABASE GUIDELINES..............................................................1 LOOKUP VS.........129 12..................................................................................................................................................................................................................2 DB2 GUIDELINES.......................................................1 WARNING ON SINGLE-NODE CONFIGURATION FILES.............................................................127 12 PERFORMANCE TUNING JOB DESIGNS...................................................1 TRANSFORMER STAGE............................................................133 12.......................................................5 INTERPRETING THE PARALLEL JOB SCORE...............................................................119 10.......................................148 APPENDIX D: ESTIMATING THE SIZE OF A PARALLEL DATA SET....................................124 11...............................................................125 11..........................................

Here is the general flow diagram for Data Stage Enterprise Edition jobs: Before Job Subroutine Halt on Error ? Yes Exit Failure No Create Reject Files (Limited) Read Input Data Halt on Error ? Yes Exit Failure No Create Error and Reject Files Perform Validations Errors and Warnings Halt on Error ? Yes Exit Failure No Create Error and Reject Files Perform Transformations Halt on Error ? Yes Exit Failure No Create Reject Files (Limited) Perform Load and/or Create Intermediate Datasets Over Job W arning Threshold? Yes Exit Failure No Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence 1 Data Integration Overview Work performed by Data Integration jobs fall into 4 general categories: • • • Reading input data including sequential files. databases and DS/EE Data Sets. . and • Provisioning data targets. 2006 6 of 179 After Job © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. transmitted. Performing transformation from data sources to data targets. transcribed. stored in a Subroutine retrieval system. Performing row validation to support data quality. All rights reserved. No part of this publication may be reproduced.

etc). These topics are discussed in Parallel Framework Standard Practices: Administration. Cron. Job Sequences require a level of integration with various production automation technologies (scheduling.Information Integration Solutions Center of Excellence 1. CA7. All rights reserved. etc). Management. This provides a level of granularity and control that is easy to manage and maintain. and together form a single end-to-end module within a DataStage application. . as illustrated in the example below: These job Sequences control the interaction and error handling between individual DataStage jobs. transcribed. Job sequences also provide the recommended level of integration with external schedulers (such as AutoSys. No part of this publication may be reproduced.1 Job Sequences As shown in the previous diagram. and provides an appropriate leveraging of the respective technologies. error logging. ETL development is intended to be modular. and Production Automation. built from individual Parallel jobs assembled in DataStage Enterprise Edition (“DS/EE”) controlled as modules from master DataStage Sequence jobs. auditing/capture. Parallel Framework Red Book: Data Flow Job Design July 17. In most production deployments. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. transmitted. 2006 7 of 179 © 2006 IBM Information Integration Solutions.

and preserves the compute effort of long running transformation jobs. No part of this publication may be reproduced. The following table defines when each type should be used: Type Transformatio n Data Requirements Data must NOT be changed by any method unless jobs transforming an entire subject area have successfully completed. This prevents partial replacement of reference data in the event of transformation failure. 2006 8 of 179 © 2006 IBM Information Integration Solutions. Transformation jobs prepare data for provisioning jobs Provisioning jobs load transformed data. All rights reserved. . Example Reference tables upon which all subsequent jobs and/or the current data target (usually a database) will depend. The data target (usually a database) must allow subsequent processing of error or reject rows and tolerate partial or complete non-update of targets. transcribed. or long running provisioning processes. and Provisioning. or translated into any language in any form by any means without the written permission of IBM. or where the resource requirements for data transformation are very large. stored in a retrieval system. Neither the transformation nor provisioning requirements are large. Hybrid Provisioning Data must NOT be changed by any method unless jobs transforming an entire subject area have successfully completed. and Hybrid jobs do both. transmitted.Information Integration Solutions Center of Excellence 1. Non-reference data or independent data are candidates. Data can be changed regardless of success or failure.2 - Job Types Nearly all data integration jobs fall into three major types: Transformation. or where the resource requirements for data provisioning are very large. Any target where either all sources have been successfully transformed or where the resources required to transform the data must be preserved in the event of a load failure or where the provisioning will take so long that it increases the probability of job failure if the job includes transformation and provisioning. Parallel Framework Red Book: Data Flow Job Design July 17. Hybrid.

.2. or translated into any language in any form by any means without the written permission of IBM. some of which may be write-through cache Data Sets. stored in a retrieval system. data sources. The following example transformation job demonstrates the use of write-through cache DS/EE Data Sets: The target table is among the inputs. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 1. transcribed.1 Transformation Jobs In transformation jobs. 2006 9 of 179 © 2006 IBM Information Integration Solutions. are processed to produce a load-ready Data Set that represents either the entire target table or new records to be appended to the target table. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. If the entire target table is held in the load-ready Data Set. that Data Set qualifies as write-through cache and may be used as source data instead of the target table. transmitted.

transcribed. . transmitted. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. 2006 10 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system.Information Integration Solutions Center of Excellence The following example transformation job does NOT produce write-through cache – its sources do NOT include the target table.

transmitted. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. it is rejected by the transformer to a reject port and the validation is not performed for those records. In this case. the validated records.2. and the write-through cache records from the last load of the target database are merged. No part of this publication may be reproduced. . If the key column is NULL. producing an ordered and linked associative table. This job also loads the target database table and creates write-through cache. This enforces a coherent view of the subject area from either cache (current state if all jobs complete successfully) or target tables (previous state if any job fails). transcribed. the cache is deleted. we replicated the Oracle structure (lower input) for each country found in the write-through cache country dataset (upper input).2 Hybrid Jobs The following example hybrid job demonstrates several interesting techniques that might be used in more complex jobs. In this case. if the load fails. The merged records are grouped and ordered before being de-duplicated to remove obsolete records. and described below following the highlighted areas from Left to Right: A column generator inserts the key column for a join and generates a single value guaranteed to never appear in the other input(s) to the join. 2006 11 of 179 © 2006 IBM Information Integration Solutions. The key column for a Referential Integrity check is validated by a Transformer stage. By specifying a full-outer join we produce a Cartesian product dataset.Information Integration Solutions Center of Excellence 1. forcing other jobs that might depend on this data to access the existing (not updated) target database table. The non-validated records. The de-duplicated records are re-grouped and ordered before calculation of the terminating keys. Some of the more interesting solutions in this job are circled.

No part of this publication may be reproduced.2. All rights reserved. stored in a retrieval system. . 2006 12 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM.3 Provisioning Jobs This example provisioning job demonstrates the straightforward approach to simple provisioning tasks. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.Information Integration Solutions Center of Excellence 1. transmitted.

. Install. These directories are configured during product installation..1 Data. /DataN /P . / /DSEngine Data File Systems /PXEngine /Data0 /Configurations /Project_A . stored in a retrieval system. and Project Directory Structures The following diagrams depict the IBM WebSphere DataStage software directory structures and the support directory structures... transmitted.. No part of this publication may be reproduced. All rights reserved. 2006 13 of 179 /Projects / /Project_A © 2006 IBM Information Integration Solutions.. or translated into any language in any form by any means without the written permission of IBM. Development standards can also make it easier to integrate external processes such as automated auditing and reporting. .. transcribed. and Data Directories /Project_Z Parallel Framework Red Book: Data Flow Job Design July 17. and to build technical and support documentation.Information Integration Solutions Center of Excellence 2 Standards Establishing consistent development standards helps to improve developer productivity and reduce ongoing maintenance costs. /Ascential /Scratch0 /ScratchN /Project_A /patches /P /DataStage /Project_Z . Install FS Install File System Scratch File Systems ..1 Directory Structures 2.. Scratch.. 2. Figure 1: RecommendedGigabyte 1 DataStage Install. File systems are highlighted in blue /Project_Z .1.

the administrator should consider creating separate file systems for each Scratch and Resource partition. Parallel Framework Red Book: Data Flow Job Design July 17. It is a bad practice to share the DataStage project file system and conductor file system with volatile files like scratch files and Parallel data set part files. No part of this publication may be reproduced. 2006 14 of 179 © 2006 IBM Information Integration Solutions. transmitted. the DataStage Administrator client creates its projects (repositories) in the Projects directory of the DataStage installation directory. . The DataStage installation creates the following two directories: $DSHOME/.Information Integration Solutions Center of Excellence By default. This best practice advocates creating subdirectories for each project for each scratch and disk partition. To scale I/O performance within DataStage. transcribed. In order to scale I/O for DataStage. Scratch is used by the EE framework for temporary files such as buffer overflow./Scratch $DSHOME/. stored in a retrieval system. because they increase the risk of filling the DataStage project file systems. Consider naming the file systems in accordance with partition numbers in your DataStage EE Configuration file. This is illustrated in the above diagram.. For this reason. it is a bad practice to create DataStage projects in the default directory. In general. as a separate file system for the Projects sub directory within the DataStage installation./Datasets The DataStage Administrator should ensure that these default directories are never used by any parallel configuration files.. projects should be installed in their own file system. consider creating separate file systems for each Scratch and Data resource partition. it is possible to create separate file systems at non-root levels. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. as disk space is typically limited in production install directories. NOTE: On some operating systems. sort memory overflow. As a standard practice.

/archive of these development phases may be present on a local file not all system. target data files.Information Integration Solutions Center of Excellence /Staging /dev /si /qa /prod /Project_A /Project_A /Project_A /Project_A /archive /archive /archive /archive . All rights reserved. /Project_Z July 17.. and production) as appropriate. No part of this publication may be reproduced. location of source data files. stored in a retrieval system. qa.. 2006 15 of 179 . data directories are implemented for each deployment phase of a job (development. system integration. transmitted. If the file system is not /archive /archive /archive shared across multiple servers. files are separated by Project name as shown below. Within each deployment directory. /Staging /dev Top-Level Directory development data tree. /Project_Z Within the separate Staging file system. error and reject files... transcribed.. /Project_A /archive /si /qa /prod Parallel Framework Red Book: Data Flow Job Design Figure 2: DataStage Staging Directories /Project_Z /Project_Z © 2006 IBM Information Integration Solutions. subdirectory created for each project location of compressed archives created by archive process of previously processed files System Integration (also known as “test”) data tree Quality Assurance data tree Production data tree .. or translated into any language in any form by any means without the written permission of IBM. ... .

Information Integration Solutions Center of Excellence 2. July 17./Project_Z publication may be reproduced. This will provide a convenient vehicle to group and manage resources used by a project. or translated into any language in any form by any means without the written permission of IBM. transmitted. 16 of 179 . another application or middle ware. or it could require scripts for example integrating with an Enterprise Scheduler. The integration can be as simple as a file system for housing source files..1. stored in a No part of this /Project_Z /Project_Z retrieval system.2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities. The directory structure will be made transparent to the DataStage application. which will enable DataStage applications to move through the life cycle without any code changes.. A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project. /Project_Z . To completely integrate all aspects of a DataStage application the directory structure that is used for integration with external entities should be defined in a way that provides a complete and separate structure in the same spirit as a DataStage project.. Environment variables are a critical portability tool. through the use of environment variables.. such as the operating system.... . 2006 . Project_Plus Project_Plus Directory Hierar /si /qa /prod /dev /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /sql /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /sql /Project_A /bin /s /d /datase /logs /param /schema /scripts /sql Figure 3: Project_Plus/sql Directory Structure Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.. All rights reserved. . transcribed.

transmitted. All rights reserved. and production) as appropriate. BuildOps. If the file system is not shared across multiple servers. for example the following diagram shows Project_Plus variables as defined in the DataStage Administrator.Information Integration Solutions Center of Excellence Within the Project_Plus hierarchy. Parallel Framework Red Book: Data Flow Job Design July 17. . transcribed. stored in a retrieval system. directories are created for each deployment phase of a job (development. qa. Project_Plus /dev /Project_A /bin /doc /src Top-Level of Directory Hierarchy development code tree subdirectory created for each project location of custom programs. No part of this publication may be reproduced.$ProjectName project files location of Orchestrate schema files location of operating system (shell) script files location of maintenance or template SQL system integration (aka “test”) code tree quality assurance code tree production code tree /datasets /logs /params /schemas /scripts /sql /si /qa /prod In support of a Project_Plus directory structure environment variable parameters should be configured. or translated into any language in any form by any means without the written permission of IBM. a copy of dsenv and copies of DSParams. not all of these development phases may be present on a local file system. 2006 17 of 179 © 2006 IBM Information Integration Solutions. system integration. utilities. and shells location of documentation for programs found in /bin subdirectory location of source code and makefiles for items found in /bin subdirectory (Note: depending on change management policies. DataStage routines.ds file) location of custom job logs and reports location of parameter files for automated program control. this directory may only be present in the /dev development code tree) location of DataSet header files (.

A similar directory structure to the Project_Plus structure could be configured and referred to as DataStage_Plus. . A consistent naming standard is essential to • maximize the speed of development • minimize the effort and cost of downstream maintenance • enable consistency across multiple teams and projects • facilitate concurrent development • maximize the quality of the developed application • increase the readability of the objects in the visual display medium • increase the understanding of components when seen in external systems.Information Integration Solutions Center of Excellence Figure 4: Project_Plus Environment Variables In some implementations. stored in a retrieval system.2 Naming Conventions As a graphical development environment. transcribed. By default. they do not facilitate ease of maintenance over time. for example all jobs are invoked with the same Script. for example in WebSphere MetaStage. 2. or translated into any language in any form by any means without the written permission of IBM. the Designer tool assigns default names based on the object type. While the default names may create a functional data flow. transmitted. No part of this publication may be reproduced. 2006 18 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. there may be external entities that are shared with other DataStage projects. or an XML extract Parallel Framework Red Book: Data Flow Job Design July 17. DataStage offers (within certain restrictions) flexibility to developers when naming various objects and components used to build a data flow. nor do they adequately document the business rules or subject areas. and the order the item is placed on the design canvas.

Where there is no sub classification required then the class word will simply refer to the object. 2006 19 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence This section presents a set of standards and guidelines to apply to developing data integration applications using DataStage Enterprise Edition. Reference (Lookup). transcribed. enabling a greater understanding of the requirements and greater control over how they are delivered. to be tuned according to needs. and so forth. stored in a retrieval system. In these cases the class word represents the subtype. the term “Standard” refers to those principles that are required. For example. three or four letter abbreviation. Subject Modifier. As an example. a Left Outer Join). the class word refers to the functions of Reading. principles. objects can be sub-typed (for example. Table. so it is envisaged that these standards will develop and will adapt over time to suit both the organization and the purpose. In some cases where appropriate. No part of this publication may be reproduced. more effort can be put into analysis and design. a transformer might be named: Data_Block_Split_Tfm As a guideline. 2. more readable names. the class word is used to identify either a type of object or the function that a particular type of object will perform. . • Since much of the development work is done using a click. the Class Word is represented as a two. transmitted. Moving or Writing data (or within a Sequence Job. while the term “Guideline” refers to recommended. or translated into any language in any form by any means without the written permission of IBM. and Class Word In the context of DataStage. Where it is a three or four letter abbreviation then it should be word capitalized. the moving of a message).2. In the case of a data store the class word will refer to the type of data store. Throughout this section. Sequential File. • There can be a much tighter link between design and development. for example: Data Set. while maintaining quality. in the case of a link object. drag and drop paradigm there is less typing involved hence the opportunity to use longer more meaningful. All rights reserved. View. There are a number of benefits from using a graphical development tool like DataStage. Parallel Framework Red Book: Data Flow Job Design July 17. Where it is a two letter abbreviation both letters should be capitalized. and many of these benefits were used to establish this naming standard: • With rapid development. but not required. Any set of standards needs to take on the culture of an organization.1 Key Attributes of the Naming Convention This naming convention is based on a three-part convention: Subject.

transmitted. This can help make them more productive and makes their work more easily read. Subject Modifier. Carrying this information as a separate attributes enables the first word of the name to be used as the subject matter. when creating the object. Secondly the class word approach enables sub-classification by object type to provide additional information. or translated into any language in any form by any means without the written permission of IBM. however. and so forth. the first letter of each Word should be capitalized. 2.2. This should be enhanced by also using Word Capitalization.2. consideration should be made to provide DataStage developers with higher resolution screens as this provides them with more screen display real-estate. expanding the icon border can significantly improve readability. Though DataStage imposes some limitations on the type of characters and length of various object names. a job design. attention should be given to the layout to enhance readability before it is handed over to versioning. The “Snap to Grid” feature of Designer can help improve development speed. When development is more or less complete. a link. allowing sort either by subject matter or by object type.2 Designer Object Layout The effective use of naming conventions means that objects need to be spaced appropriately on the DataStage Designer canvas. stored in a retrieval system. DataStage provides the ability to document during development with the use of meaningful naming standards (as outlined in this section). . One benefit of using the Subject. whatever tool you use. the object type is defined in a separate field. 2006 20 of 179 © 2006 IBM Information Integration Solutions. so a pattern of work needs to be identified and adopted to help development.there is a field that denotes whether the object is a column. This is the same or similar information that would be carried in a prefix approach. For stages with multiple links. No part of this publication may be reproduced. documentation is often something that is left until later. the abbreviated form is used. Though best intentions are always apparent. all word abbreviations should be referenced by the long form to get used to saying the name in full even if reading the abbreviation. 2. This type of approach takes extra effort at first. a stage.Information Integration Solutions Center of Excellence A list of frequently-used Class Word abbreviations is provided in 12. Like a logical name. For the purposes of documentation. Class Word approach. Establishing standards also eases use of external tools and Parallel Framework Red Book: Data Flow Job Design July 17. over using the Prefix approach. This will help re-enforce wider understanding of the subjects. the standard. for example.2 DataStage Naming Reference. inadequately carried out.3 Documentation and Metadata Capture One of the major problems with any development effort. In WebSphere MetaStage. Where possible.4. is to enable two levels of sorting or grouping. transcribed. is maintaining documentation. will be to separate words by an Underscore which will allow clear identification of each work in a name. a derivation. where possible. All rights reserved. The key issue is readability.

2. Test (“Test”). Therefore Directory names should be Word Capitalized and separated by either an underscore or a space. Routines.4. Table definitions…). it can contain alpha-numeric characters and it can contain underscores.4. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17. This factor often can cause terminology issues especially in teamwork where both business and developers are involved. Examples of Project naming where the project is single application focused are: • “Accounting Engine NAB Development” would be named: Acct_Eng_NAB_Dev • “Accounting Engine NAB Production” would be named: Acct_Eng_NAB_Prod Examples of Project naming where the project is multi-application focused are: • Accounting Engine Development or Acct_Engine_Dev • Accounting Engine Production or Acct_Engine_Prod 2. as well as documentation and auditing. dialog box fields that specify a new category have only one input area for defining the Category name. However with the limit of 18 characters the name is most often composed of abbreviations. Shared Containers. For example.3 Job Category Naming Within Designer.2.2. stored in a retrieval system.2 Category Hierarchy DataStage organizes objects in its repository by Categories. developers have the flexibility to define their own Directory or Category hierarchy.2. transcribed. 2. All rights reserved. The suffix of a Project name should be used to identify Development (“Dev”). Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”). Jobs. No part of this publication may be reproduced.1 Projects Each DataStage Project is a standalone repository. are Alpha Numeric and can also contain both Spaces and Underscores. allowing related objects to be grouped together. It may or may not have a one to one relationship with an organizations’ project of work. and Production (“Prod”). or translated into any language in any form by any means without the written permission of IBM. .4 Naming Conventions by Object Type 2. The name of a DataStage Project may only be 18 characters in length. Category Names can be long. DataStage enforces the top level Directory Structure for different types of Objects (for example.4.Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage. 2006 21 of 179 © 2006 IBM Information Integration Solutions. which can provide impact analysis. Below this level. transmitted. 2.

stored in a retrieval system. a Category level should only contain objects that are directly related. Categorization by Functional Module For a given application.Information Integration Solutions Center of Excellence Figure 5: Creating Category Hierarchies The main reason for having Categories is to group related objects. with sub-levels for individual functional modules. Note that Job names must be unique within a DataStage project. a job category might contain a Job Sequence and all the jobs and only those jobs that are contained in that sequence. For example. Organizing related DataStage objects within categories also facilitates backup/export/import/change control strategies for projects since Manager can import/export objects by category grouping. all Jobs and Job Sequences will be grouped in a single parent Category. not within a category. jobs that read write-through cache for a ECRP subset in the ECRDEV project that cleanse and load multi-family mortgage data and are driven by a sequencer might have a hierarchy that looks like the following example: Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. For example. . Where possible. transcribed. transmitted. Within each functional module category. 2006 22 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. Jobs and Job Sequences are grouped together in the same scope as the technical design documents.

and the responsibility of the development manager assigned the DataStage Manager role to ensure that projects are not obese with unused jobs. When implementing a customized Table Definition categorization. On import. No part of this publication may be reproduced.. or translated into any language in any form by any means without the written permission of IBM. Saved.Information Integration Solutions Center of Excellence Figure 6: Categorization by Functional Module Categorization by Developer In development projects. Table Definitions are always categorized using two level names. Temporary TableDefs created by developers to assist with job creation appear under the Saved category by default. Remembering that Job names must be unique within a given project. Once created. ECRP and Templates. It is the responsibility of each developer to delete unused or obsolete code. All rights reserved. categories and metadata. establishing a Table Definition categorization that matches project development organization is recommended. 2.2. 2006 23 of 179 © 2006 IBM Information Integration Solutions. New Table Definition categories can be created within the repository by right-clicking within the Table Definitions area of the DataStage project repository and choosing the “New Category” command. categories will be created for each developer as their personal sandbox and place they perform unit test activities on jobs they are developing. In the previous illustration. two developers have private categories for sandbox and development activities. . if these TableDefs are to be used by other jobs. Orchestrate. The placement of these fields varies with the method of metadata import.4 Table Definition Categories Unlike other types of DataStage objects.). PlugIn. transmitted. project manager. but this can be overridden during import. they must be moved to the appropriate category and re-imported from that category in every job where they are used. and there are 2 additional high-level categories. By default. care must be taken to override the default choices for category names during Table Definition import. Although the default table definition categories are useful from a functional perspective. transcribed. stored in a retrieval system. the first level Table Definition category is identified as the “Data Source Type” and the second level categorization is referred to as the “Data Source Name” as shown in the example on the below.. etc. two developers cannot save a copy of the same job with the same name within their individual “sandbox” categories – a unique Job name must be given. TableDefs that remain in the Saved category Parallel Framework Red Book: Data Flow Job Design July 17. DataStage assigns the level names based on the source of the metadata import (for example.4.

: Datasets. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”. In this example.2 Category Hierarchy. Each subject area will have a master category. This will be discussed further in Section 2. . and underscores only.g: DWPH1 or ECRP. the TableDefs have been grouped into a master category of Custom. Examples of Job naming are: • CodeBlockAggregationJob • CodeBlockProcessingSeq Jobs should be organized under Category Directories to provide grouping such that a Directory should contain a Sequence Job and all the Jobs that are contained within that sequence.2. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”. stored in a retrieval system.2. Because the name of can be long. 2006 24 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence should be deleted as soon as possible. with sub-categories intended to identify the type of the source. numbers. e. Figure 7: Table Definition Categories 2. e.4.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category. transmitted. Job and Job Sequence names should be descriptive and should use word capitalization to make them readable. All rights reserved.g.4. or translated into any language in any form by any means without the written permission of IBM. An alternative implementation is to set the “Data source name” to that of the source system or schema. No part of this publication may be reproduced.

instead of using the default “DSLink#” (where “#” is an assigned number). 2006 25 of 179 © 2006 IBM Information Integration Solutions. It is particularly important to establish a consistent naming convention for link names.8 Links Within a DataStage Job. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17. Within a Job Sequence.6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. The class word suffix is “Parm”.7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores.4. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. transcribed.4.Information Integration Solutions Center of Excellence 2. a character code is automatically added to that instance of its use throughout the project. so word capitalization should be used for readability. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2. Within the graphical Designer environment.2. links are objects that represent the flow of data from one stage to the next.2. . All rights reserved. during Link Ordering). It is optional as to whether you decide to change this code to something meaningful. Examples of Shared Container naming are: • AuditTrailPsc (this is the original as seen in the Category Directory) • AuditTrailPscC1 (This is an instance of use of the above shared container) • AuditTrailPscC2 (This is another instance of use of the same shared container) In the above examples the characters “C1” and the “C2” are automatically applied to the Shared Container Stage by DataStage Designer when dragged onto the design canvas. When a Shared Container is used. stored in a retrieval system. Furthermore.2. the following Class Word naming is recommended: • Psc = Parallel (Enterprise Edition) Shared Container • Ssc = Server Edition Shared Container IMPORTANT: Use of Server Edition Shared Containers is discouraged within a parallel job. stage editors identify links by name. links represent the flow of a message from one activity / step to the next. having a descriptive link name reduces the chance for errors (for example. transmitted. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. To differentiate between Parallel Shared Containers and Server Shared Containers. 2.4. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. Therefore the parameter name must be made readable using Capitalized words separated by underscores.

Parallel Framework Red Book: Data Flow Job Design July 17. a 2. In this case.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. or 4 character abbreviation should be used for the Class Word suffix.4. always specify a unique link name within a particular Job or Job Sequence by including a number. stored in a retrieval system. These names are based on the type of stage (object) and a unique number. Sequential File. establishing standardized link names makes it easier to understand results and audit counts.4. for example: o “In” for input o “Out” for output o “Upd” for updates o “Ins” for inserts o “Del” for deletes o “Get” for shared container inputs o “Put” for shared container output As data is enriched through stages. . All rights reserved. Within a Job or Job Sequence. No part of this publication may be reproduced.2 DataStage Naming Reference. after the subject name and subject modifier. Merge. etc) • The type of movement may optionally be part of the Class Word. reject) to reinforce the visual cues of the Designer canvas: o “Ref” for reference links (Lookup) o “Rej” for reject links (Lookup. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12. based on the order the object was added to the flow. (The DataStage Designer does not require link names on different stages to be unique. Transformer. the same name may be appropriate for multiple links. 2006 26 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence example. transcribed.2. The following rules can be used to establish a link name: • Use the prefix “lnk_” before the subject name to differentiate with stage objects • The link name should define the subject of the data that is being moved • For non-stream links. through Job reporting). stage names must be unique. transmitted. Database. or translated into any language in any form by any means without the written permission of IBM.) • Examples Link names: • Input Transactions: “lnk_Txn_In” • Reference Account Numbers: “lnk_Account_Ref” • Customer File Rejects: “lnk_Customer_Rej” • Reception Succeeded Message or “lnk_Reception_Succeeded_Msg” 2. Instead of using the full object name. the link name should include the link type (reference. 3.

Information Integration Solutions Center of Excellence

2.2.4.10 Data Stores For the purposes of this section, a data store is a physical piece of disk storage where data is held for some period of time. In DataStage terms, this can be either a table in a database structure or a file contained within a disk directory or catalog structure. Data held in a database structure is referred to as either a Table or a View. In data warehousing, two additional subclasses of table might be used: Dimension and Fact. Data held in a file in a directory structure will be classified according to its type, for example: Sequential File, Parallel Data Set, Lookup File Set, etc. The concept of source and target can be applied in a couple of ways. Every job in a series of jobs could consider the data it gets in to be a source and the data it writes out as being a target. However for the sake of this naming convention a Source will only be data that is extracted from an original system and Target will be the data structures that are produced or loaded as the final result of a particular series of jobs. This is based on the purpose of the project – to move some data from a source to a target. Data Stores used as temporary structures to land data between jobs, supporting restart and modularity, should use the same names in the originating job and any downstream jobs reading the structure. Examples of Data Store naming are: • Transaction Header Sequential File or Txn_Header_SF • Customer Dimension or Cust_Dim (This optionally could be further qualified as Cust_Dim_Tgt if you wish to qualify it as a final target) • Customer Table or Cust_Tab • General Ledger Account Number View or GL_Acctno_View 2.2.4.11 Transformer Stage and Stage Variables A Transformer Stage name can be long – over 50 characters and can contain underscores. Therefore the name can be descriptive and readable through word capitalization and underscores. DataStage Enterprise Edition supports two types of Transformers: • “Tfm”: Parallel (Enterprise Edition) Transformer • “BTfm”: BASIC (Server Edition) Transformer IMPORTANT: For maximum performance and scalability, BASIC Transformers should be avoided in Enterprise Edition data flows. A Transformer Stage Variable can have a long name consisting of alphanumeric characters but not underscores. Therefore the Stage Variable name must be made readable only by using Capitalized words. The Class Word suffix is Stage Variable or “SV”. Stage Variables should be named according to their purpose.

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

27 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

When developing Transformer derivation expressions, it is important to remember Stage variable names are case sensitive. 2.2.4.12 DataStage Routines DataStage BASIC routine names will indicate their function and they will be grouped in sub-categories by function under a main category of Custom, for example.: Routines/Custom/SetDSParamsFromFile. A How-To document describing the appropriate use of the routine must be provided by the author of the routine, and placed in a documentation repository. DataStage Custom Transformer routine names will indicate their function and they will be grouped in sub-categories by function under a main category of Custom, for example: Routines/Custom/DetectTeradataUnicode. Source code, a makefile, and the resulting object for each Custom Transformer routine must be placed in the project phase source directory, e.g.: /home/dsadm/dev/bin/source. 2.2.4.13 File Names Source file names should include the name of the source database or system and the source table name or copybook name. The goal is to connect the name of the file with the name of the storage object on the source system. Source flat files will have a unique serial number composed of the date, “_ETL_” and time, for example: Client_Relationship_File1_In_20060104_ETL_184325.psv. Intermediate datasets are created between modules. Their names will include the name of the module that created the dataset OR the contents of the dataset in that more than one module may use the dataset after it is written, for example: BUSN_RCR_CUST.ds Target output files will include the name of the target database or system, the target table name or copybook name. The goal is the same as with source files – to connect the name of the file with the name of the file on the target system. Target flat files will have a unique serial number composed of the date, “_ETL_” and time, for example: Client_Relationship_File1_Out_20060104_ETL_184325.psv Files and datasets will have suffixes that allow easy identification of the content and type. DataStage proprietary format files have required suffixes and are identified in italics in the table below which defines the types of files and their suffixes. File Type
Parallel Framework Red Book: Data Flow Job Design

Suffix
July 17, 2006 28 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Flat delimited and non-delimited files Flat pipe (|) delimited files Flat comma-and-quote delimited files DataStage datasets DataStage filesets DataStage hash files Orchestrate schema files Flat delimited or non-delimited REJECT files DataStage REJECT datasets Flat delimited or non-delimited ERROR files DataStage ERROR datasets Flat delimited or non-delimited LOG files .dat. .psv .csv. .ds. .fs .hash. .schema. .rej. _rej.ds. .err. _err.ds. .log.

2.3

Documentation and Annotation

DataStage Designer provides description fields for each object type. These fields allow the developer to provide additional descriptions that can be captured and used by administrators and other developers. The Short Description field is also displayed on summary lines within the Director and Manager clients. At a minimum, description annotations must be provided in the Job Properties Short Description field for each job and job sequence, as shown below:

Figure 8: Job Level Short Description Within a job, the Annotation tool should be used to highlight steps in a given job flow. Note that by changing the vertical alignment properties (for example, Bottom) the annotation can be drawn around the referenced stage(s), as shown in the following example.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 29 of 179
© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

stored in a retrieval system. There are no selection criteria in the WHERE clause. There are no selection criteria in the WHERE clause.Information Integration Solutions Center of Excellence Figure 9: Example Job Annotation DataStage also allows descriptions to be attached to each stage within the General tab of the stage properties. These descriptions will appear in the job documentation automatically generated from jobs and sequencers adhering to the standards in this document. All rights reserved. . 2006 30 of 179 © 2006 IBM Information Integration Solutions. ODBC Enterprise stage read: Read the GLO.RcR_GLOBAL_BUSN_CAT_TYP table from jpORACLE_SERVER using the ODBC driver. Oracle Enterprise stage read: Read the GLOBAL. transmitted.GLOBAL_REST_CHAR table from jpORACLE_SERVER using the Oracle Enterprise operator. Each stage should have a short description of its function specified within the stage properties. Examples of such annotations include: Job “short” description: This Job takes the data from GBL Oracle Table AD_TYP and does a truncate load into Teradata Table AD_TYP. or translated into any language in any form by any means without the written permission of IBM. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. More complex operators or operations should have correspondingly longer and more complex explanations on this tab. No part of this publication may be reproduced.

This stage converts null dates. No part of this publication may be reproduced. Copy stage This stage sends data to the TDMLoadPX stage for loading into Teradata. or translated into any language in any form by any means without the written permission of IBM. environment files. Sequential file stage: This is the source file for the LANG table. and to a dataset for use as write-through cache. etc. 2006 31 of 179 © 2006 IBM Information Integration Solutions. Lookup stage This stage validates the input and writes rejects. UNIX shell scripts. . This is the target file for business qualification process rejects. transmitted. Teradata MultiLoad stage: Load the RcR_GLOBAL_LCAT_TYP table. However. sequences. Parallel Framework Red Book: Data Flow Job Design July 17. Modify stage: This stage performs data conversions not requiring a transformer. Transformer stage: This stage generates sequence numbers that have a less-than file scope.) that may be part of a completed application. This stage validates the input and continues. This stage renames and/or drops columns and is NOT optimized out. This stage identifies changes and drops records not matched (not updated).4 Working with Source Code Control Systems DataStage’s built-in repository manages objects (jobs. which is used as write-through cache to avoid the use of Teradata. Data Set stage: This stage writes the GLOBAL_Ad_Typ dataset which is used as write-through cache to avoid the use of Teradata in subsequent jobs. custom components) during job development.Information Integration Solutions Center of Excellence Remove Duplicates stage This stage removes all but one record with duplicate BUSN_OWN_TYP_ID keys. This stage is cosmetic and is optimized out. this repository is not capable of managing nonDataStage components (for example. job scheduler configurations. All rights reserved. stored in a retrieval system. routines. This stage reads the GLOBAL_Lcat dataset. 2. transcribed. table definitions.

2006 32 of 179 © 2006 IBM Information Integration Solutions. it is important that an identified individual maintains backup copies of the important job designs using . It would be cumbersome to require the developer to check-in every change to a DataStage object in the external source code control system.XML. DataStage does not directly integrate with source code control systems. etc. you can export objects (job designs. Rather. rules should be defined for when this transfer should take place. Grouping related objects by folder can simplify the process of exchanging information with the external source code control system. It is the responsibility of the DataStage developer to maintain DataStage objects within the source code system. Unless there is a need to parse information in the export file. user-defined routines. Furthermore. . For these reasons. The export file format for DataStage objects can be either . transmitted.for example. SCCS) are useful for managing the development lifecycle of all components of an application. 2. stored in a retrieval system. These files can then be checked into the external source code control system. custom stage types.DSX file exports to a local or (preferably) shared file system. milestone points on the development lifecycle are a good point for transferring objects to the source code control system . and cannot be used to restore individual objects. Using Manager. and so on. they would not be an effective backup strategy. Typically. Designer facilitates iterative job design. operating system backups of the project repository files only establish a “point in time”. b) Define rules for exchange with source code control As a graphical development environment. table definitions. when a set of objects has completed initial development. although the XML file is generally much larger. unit test.Information Integration Solutions Center of Excellence Source code control systems (such as ClearCase. but it does offer the ability to exchange information with these systems. This object grouping also helps establish a manageable “middle ground” between an entire project exports and individual object exports. organized into specific releases for version control.) from the repository as clear-text format files. transcribed. . or translated into any language in any form by any means without the written permission of IBM. c) Don’t rely on the source code control system for backups Because the rules defined for transfer to the source code control system will typically be only at milestones in the development cycle. All rights reserved.4. Both formats contain the same information.DSX (DataStage eXport format) or . The Manager client is the primary interface to the DataStage object repository.1 Source Code Control Standards The first step to effective integration with source code control systems is to establish standards and rules for managing this process: a) Establish Category naming and organization standard DataStage objects can be exported individually or by category (folder hierarchy). No part of this publication may be reproduced.DSX is the recommended export format. PVCS. Parallel Framework Red Book: Data Flow Job Design July 17.

transcribed. All rights reserved. 2. or translated into any language in any form by any means without the written permission of IBM. . Assigning related objects to the same category provides a balanced level of granularity when exporting and importing objects with external source code control systems. All exports from the DataStage repository are performed on the Windows workstation. There is no server-side project export facility. the developer should create a local backup prior to implementing any extensive changes.2 Category Hierarchy. establishing and following a consistent naming and categorization standard is essential to the change management process. July 17. as explained in this section. In either case. or by the individual DataStage developer.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process. The DataStage client includes Windows command-line utilities for automating the export process. Category.4. These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide. and individual Object levels. The DataStage Manager can export at the Project. It can be done interactively by the developer or project manager using the Manager client. stored in a retrieval system. • Select the object or category in the Manager browser.4. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group. 2.2. transmitted.4. No part of this publication may be reproduced.2 Using Object Categorization Standards As discussed in Section 2. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu.

check-in the exported . 2006 34 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.4 Import from Source Code Control System In a similar manner.Information Integration Solutions Center of Excellence NOTE: Objects cannot be exported from DataStage if they are open in Designer. the import of objects from an external source code control system is a straightforward process. stored in a retrieval system. transcribed. . If you wish to include compiled Transformer objects for a selected job.4.DSX file 2. Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. the option “Selection: By category” should be specified in the “Options” tab. No part of this publication may be reproduced. Make sure all objects are saved and closed before exporting. The filename for export is specified in the “Export to file:” field at the top of the Export dialog. • Using your source code control utilities. Import can be interactive through the Manager client (as described in this section). or translated into any language in any form by any means without the written permission of IBM. Figure 11: Manager Export Options To export a group of objects to a single export file. make sure the “Job Executables” category is checked. or automated through command-line utilities.

DSX file using Manager. Choose “Import DataStage Components” from the “Import” menu. The Windows workstation utilities (dsimport and dscmdimport) are documented in the DataStage Manager Guide. stored in a retrieval system. Figure 12: Manager Import options • • The import of the . Note that using dsjob will only import job executables job designs can only be imported using the Manager client or the dsimport or dscmdimport client tools. click OK to import.5 Understanding a Job’s Environment DataStage Enterprise Edition provides a number of environment variables to enable / disable product features and to fine-tune job performance. it is possible to import the job executables from the DataStage server host using the dsjob command-line. No part of this publication may be reproduced. there is a defined order of precedence that is evaluated when a job’s actual environment is established at runtime: 1) The daemon for managing client connections to the DataStage server engine is called dsrpcd. If the objects were not exported with the “Job Executables”. Although operating system environment variables can be set in multiple places. Select the file you checked out of your source code control system by clicking on the ellipsis (“…”) next to the filename field in the import dialog. • • Use the source code control system to check-out (or export) the . 2006 35 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. After selecting your file. . Import objects in the . transmitted. and should start Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence Unlike the export process. as documented in the DataStage Development Kit chapter of the Parallel Job Advanced Developer’s Guide. 2. transcribed.DSX file will place the object in the same DataStage category it originated from. dsrpcd is started when the server installed. command-line import utilities are available for both Windows workstation and DataStage server platforms. By default (in a root installation). then compile the imported objects from Designer. This means that if necessary it will create the Job Category if it doesn't already exits. For test and production environments. or translated into any language in any form by any means without the written permission of IBM.DSX file to your client workstation. or using the Multi-Job Compile tool.

the dsrpc environment is not inherited since DataStage jobs do not execute on the conductor node. . dsrpcd can also be manually started and stopped using the $DSHOME/uv – admin command. All rights reserved. Note that client connections DO NOT pick up per-user environment settings from their $HOME/. the Designer parses the corresponding parallel configuration file to obtain a list of node maps and constraints (advanced stage properties). which.Information Integration Solutions Center of Excellence whenever the machine is restarted. These settings are stored in a file named DSPARAMS in the project directory. see the DataStage Administrator Guide. Typically this is used to pickup values set in the operating system outside of DataStage. 2) Environment variable settings for particular projects can be set in the DataStage Administrator client. $ENV NOTE: $ENV should not be used for specifying the default $APT_CONFIG_FILE value because. transmitted. stored in a retrieval system. No part of this publication may be reproduced. To avoid hard-coding default values for job parameters. or by carefully editing the DSPARAMS file within the project. Any project-level settings for a specific environment variable will override any settings inherited from dsrpcd. transcribed. $PROJDEF Parallel Framework Red Book: Data Flow Job Design July 17. the default DataStage environment is defined in the registry. 2006 36 of 179 © 2006 IBM Information Integration Solutions.profile script. IMPORTANT: When migrating projects between machines or environments. or translated into any language in any form by any means without the written permission of IBM. it is important to note that project-level environment variable settings are not exported when a project is exported.) By default. there are three special values that can be used for environment variables within job parameters: • causes the value of the named environment variable to be retrieved from the operating system of the job environment. DataStage jobs inherit the dsrpcd environment. 3) Within Designer. • causes the project default value for the environment variable (as shown on the Administrator client) to be picked up and used to set the environment variable and job parameter for the job. on UNIX platforms is set in the etc/profile and $DSHOME/dsenv scripts. Any project-level environment variables must be set for new projects using the Administrator client. (For more information. On Windows. during job development. Management. environment variables may be defined for a particular job using the Job Properties dialog box. Any job-level settings for a specific environment variable will override any settings inherited from dsrpcd or from project-level defaults. On USS environments. Refer to the DataStage Administration. and Production Automation Best Practice for additional details.

2 Understanding the Parallel Job Score) Includes a copy of the generated osh in the job’s DataStage log Outputs record counts to the DataStage job log as each operator completes processing. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. providing detailed information about actual job flow including operators. (see 12.4. Extremely useful for understanding how a job actually ran in the environment. All rights reserved. This setting should be disabled by default. transcribed.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. or translated into any language in any form by any means without the written permission of IBM. Outputs EE score dump to the DataStage job log.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. Maximum buffer delay in seconds $OSH_PRINT_SCHEMA 0 $APT_PM_SHOW_PIDS 1 $APT_BUFFER_MAXIMUM_TIMEOUT 1 Parallel Framework Red Book: Data Flow Job Design July 17.5. This setting should be disabled by default. APT_SORT_INSERTION_CHECK_ONLY). . This section is intended to call attention to some specific environment variables. and to document a few that are not part of the documentation.1. 2. or may be set on an individual basis within the properties for each job.5. The count is per operator per partition. stored in a retrieval system. 2006 37 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. Does not report PIDs of DataStage “phantom” processes started by Server shared containers. Several environment variables are evaluated only for their presence in the environment (for example. $UNSET 2. processes. This variable should be included in all job parameters so that it can be easily changed at runtime. and Data Sets. Outputs actual runtime metadata (schema) to DataStage job log. It may be helpful to create a Job Template and include these environment variables in the parameter settings. Environment Variable $APT_CONFIG_FILE Setting filepath 1 $APT_DUMP_SCORE $OSH_ECHO $APT_RECORD_COUNTS 1 0 Description Specifies the full pathname to the EE configuration file. transmitted. but part of every job design so that it can be easily enabled for debugging purposes. These settings can be made at the project level. but part of every job design so that it can be easily enabled for debugging purposes.

the environment variable $APT_PM_NO_SHARED_MEMORY should be set to 1 to work around a performance issue with shared memory MMAP operations. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17. .1A platforms only: On Tru64 platforms. transcribed. transmitted. This setting instructs EE to use named pipes rather than shared memory for local data transport.1. stored in a retrieval system.5. 2. The environment variables mentioned in this document are summarized in Appendix D: Environment Variable Reference. you must define the environment variable $APT_IO_NOMAP On Tru64 5. 2006 38 of 179 © 2006 IBM Information Integration Solutions. assisting in debugging.Information Integration Solutions Center of Excellence On Solaris platforms only: When working with very large parallel Data Sets (where the individual data segment files are larger than 2GB).2 Additional Environment Variable Settings Throughout this document. or translated into any language in any form by any means without the written permission of IBM. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. All rights reserved. or changing the default behavior of specific Enterprise Edition stages. a number of environment variables will be mentioned for tuning the performance of a particular job flow.

For example. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. . It is for these reasons that long-running tasks are often segmented into separate jobs in an overall sequence. etc) and directories where files are stored. All rights reserved. • • • 3. Without these checkpoints.1 Modular Development Modular development techniques should be used to maximize re-use of DataStage jobs and components: • Job parameterization allows a single job design to process similar logic instead of creating multiple copies of the same job. To ease re-use. password. user. intermediate work files. if the extract of source data takes a long time (such as an FTP transfer over a wide area network) it would be good to land the extracted source data to a parallel data set before processing. processing must be restarted from the beginning of the job flow. it may be appropriate to update all dimension values before inserting new entries in a data warehousing fact table. Create a standard directory structure outside of the DataStage project directory for source and target files. o As another example. these standard parameters and settings should be made part of a Designer Job Template. Where possible. functional requirements may dictate job boundaries. complex job that satisfies given functional requirements. A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN. But functional requirements may not be the only factor driving the size of a given DataStage job. No part of this publication may be reproduced. In some cases. While it may be possible to construct a large. create re-usable components such as parallel shared containers to encapsulate frequently-used logic.2 Establishing Job Boundaries It is important to establish appropriate job boundaries when developing with DS/EE. 2006 39 of 179 © 2006 IBM Information Integration Solutions. Factors to consider when establishing job boundaries include:  Establishing job boundaries through intermediate Data Sets creates “checkpoints” that can be used in the event of a failure when processing must be restarted. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. transcribed. o For example.Information Integration Solutions Center of Excellence 3 Development Guidelines 3. it is generally a good idea to land data to a parallel Data Set before loading to a target database unless the data volume is small or the overall time to process the data is minimal. this may not be appropriate. and so forth. or translated into any language in any form by any means without the written permission of IBM. transmitted.

Information Integration Solutions Center of Excellence

Larger, more complex jobs require more system resources (CPU, memory, swap) than a series of smaller jobs, sequenced together through intermediate Data Sets. Resource requirements are further increased when running with a greater degree of parallelism specified by a given configuration file. However, the sequence of smaller jobs generally requires more disk space to hold intermediate data, and the speed of the I/O subsystem can impact overall end-to-end throughput. Section 12.3: Minimizing Runtime Processes and Resource Requirements provides some recommendations for minimizing resource requirements of a given job design, especially when the volume of data does not dictate parallel processing.

Breaking large job flows into smaller jobs may further facilitate modular development and reuse if business requirements for more than one process depend on intermediate data created by an earlier job. The size of a job directly impacts the speed of development tasks such as opening, saving, and compiling. These factors may be amplified when developing across a wide-area or high-latency network connection. In extreme circumstances this can significantly impact developer productivity and ongoing maintenance costs. The startup time of a given job is directly related to the number of stages and links in the job flow. Larger more complex jobs require more time to startup before actual data processing can begin. Job startup time is further impacted by the degree of parallelism specified by the parallel configuration file. Remember that the number of stages in a parallel job includes the number of stages within each shared container used in a particular job flow.

As a rule of thumb, keeping job designs to less than 50 stages may be a good starting point. But this is not a hard-and-fast rule. The proper job boundaries are ultimately dictated by functional / restart / performance requirements, expected throughput and data volumes, degree of parallelism, number of simultaneous jobs and their corresponding complexity, and the capacity and capabilities of the target hardware environment. Combining or splitting jobs is relatively easy, so don't be afraid to experiment and see what works best for your jobs in your environment.

3.3

Job Design Templates

DataStage Designer provides the developer with re-usable Job Templates, which can be created from an existing Parallel Job or Job Sequence using the “New Template from Job” command.

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

40 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Template jobs should be created with: - standard parameters (for example, source and target file paths, database login properties…) - environment variables and their default settings (as outlined in Section 2.5.1 Environment Variable Settings) - annotation blocks In addition, template jobs may contain any number of stages and pre-built logic, allowing multiple templates to be created for different types of “standardized” processing. By default, the Designer client stores all job templates in the local “Templates” directory within the DataStage client install directory, for example, C:\Program Files\Ascential\DataStage751\Templates To facilitate greater re-use of job templates, especially in a team-based development, the template directory can be changed using the Windows Registry Editor. This change must be made on each client workstation, by altering the following registry key:
HKEY_LOCAL_MACHINE\SOFTWARE\Ascential Software\DataStage Client\CurrentVersion\Intelligent Assistant\Templates

3.4

Default Job Design

Default job designs include all of the capabilities detailed Section 2: Standards. Template jobs should contain all the default characteristics and parameters the project requires. These defaults provide at a minimum: 1. 2. 3. 4. 5. Development phase neutral storage (e.g.: dev, si, qa and prod); Support for Teradata, Oracle, DB2/UDB and SQL Server login requirements; Enforced project standards; Optional operational metadata (runtime statistics) suitable for loading into a database; and Optional auditing capabilities.

The default job design specifically will support the creation of write-through cache in which data in load-ready format is stored in DS/EE Data Sets for use in the load process or in the event the target table becomes unavailable. The default job design incorporates several features and components of DataStage that are used together to support tactical and strategic job deployment. These features include: 1. Re-start-able job sequencers which manage one or more jobs, detect and report failure conditions, provide monitoring and alert capabilities and support checkpoint restart functionality. 2. Custom routines written in DataStage BASIC (DS Basic) that detect external events, manage and manipulate external resources, provide enhanced notification and alert capabilities and interface to the UNIX operating system. 3. DataStage Enterprise Edition (DS/EE) ETL jobs that exploit job parameterization, runtime UNIX environment variables, and conditional execution.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 41 of 179
© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Each subject area is broken into sub-areas and each sub-area may be further subdivided. These sub-areas are populated by a DataStage job sequencer utilizing 2 types of DataStage jobs at a minimum: 1. A job that reads source data and • Transforms it to load-ready format • Optionally stores its results in a write-through cache DataStage Data Set or loads the data to the target table. 2. A job that reads the DataStage dataset and loads it to the target table. Other sections will discuss in detail each of the components and give examples of their use in a working example job sequencer.

3.5

Job Parameters

Parameters are passed to a job as either DataStage job parameters or as environment variables. Job parameters can be set from a file and are distinguished by the presence of a ‘jp’ prefix to the variable name. This prefix is part of the DataStage development standard. The names of environment variables have no prefix when they are set (UNIX_VAR=”some value”) and a prefix of “$” when used (myval=$UNIX_VAR). Job parameters are passed from a job sequencer to the jobs in its control as if a user were answering the runtime dialog questions displayed in the DataStage Director job-run dialogue. Default environment variables cannot be reset during this dialog unless explicitly specified in the job. The scope of a parameter depends on their type. Essentially: o The scope of a job parameter is specific to the job in which it is defined and used. Job parameters are stored internally within DataStage for the duration of the job, and are not accessible outside that job. o The scope of a job parameter can be extended by the use of job sequencer, which can manage and pass job parameters among jobs. o The scope of an environment variable is wider, as it is defined at operating system level, though conversely the use of environment variables is limited within this exercise. Job parameters are required for the following DataStage programming elements: 1. File name entries in stages that use files or Data Sets must NEVER use a hard-coded operating system pathname. a. Staging area files must ALWAYS have pathnames as follows: /jpSTAGING/jpENVIRON/jpSUBJECT_AREA[filename.suffix] b. DataStage datasets ALWAYS have pathnames as follows: /jpDSTAGE_ROOT/jpENVIRON/datasets/[filename.suffix]

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

42 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

data rule violations or other reasons that would disqualify a row from processing. userid and password. Management. Because Parallel Shared Containers are inserted when a job is compiled. For maximum component re-use. The exact policy for each reject is specified in the job design document. Parallel Framework Red Book: Data Flow Job Design July 17. The presence of rejects and errors will be detected and notification sent by email to selected staff. 3. Both rejects and errors will be archived and placed in a special directory for evaluation or other action by support staff.6 Parallel Shared Containers Parallel Shared Containers allow common logic to be shared across multiple jobs. 3. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. or translated into any language in any form by any means without the written permission of IBM. Error files will include those records from sources that fail quality tests. 2006 43 of 179 © 2006 IBM Information Integration Solutions. Using RCP. All rights reserved. any additional columns are passed through the container at runtime without the need to separate and remerge. These activities are the responsibility of job sequencers used to group jobs by some reasonable grain or by a federated scheduler. Use and management of job parameters. . transmitted. transcribed. The presence of rejects may indicate that a job has failed and prevent further processing. Error rows are those rows caused by unforeseen data events such as values too large for a column or text in an unsupported language. stored in a retrieval system. and Production Automation. all jobs that use a shared container must be recompiled when the container is changed. and further. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. No part of this publication may be reproduced. as well as standardized routines for use in Job Sequencers are discussed further in Parallel Framework Standard Practices: Administration. enable RCP at the project level and for every stage within the parallel shared container. Database stages must ALWAYS use variables for the server name. The Usage Analysis and Multi-Job Compile tools can be used to recompile jobs that use a shared container.Information Integration Solutions Center of Excellence 2. The presence of errors may not prevent further processing. schema (if appropriate). This allows the container input and output links to contain only the columns relevant to the container processing. Reject files will include those records rejected from the ETL stream due to Referential Integrity failures.7 Error and Reject Record Handling Reject rows are those rows that fail active or passive business rule driven validation as specified in the job design document. whether the job or ETL processing is to continue is specified on a per-job and/or per-sequence and/or per-script basis based on business requirements.

Ignore – some process or event downstream of the ETL process is responsible for handling the error. Parallel Framework Red Book: Data Flow Job Design July 17.rej file. or translated into any language in any form by any means without the written permission of IBM.7. If a file is created by this option. it must have a *. Rejects should not exist and should stop the job. 2 3 Rejects should not exist but should not stop the job. Pass successful reads to the output stream. Only records that match the given table definition and format are output.Information Integration Solutions Center of Excellence ETL actions to be taken for each record type is specified for each stage in the job design document. transmitted. Reprocess – rows are reprocessed and re-enter the data stream. rows that cannot be read are dropped by the Sequential File stage. which outputs rows that do not match the given table definition and format specifications. Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Sequential File Stage Option Use the Continue option. a shared container error handler can be used. Rejects are tracked by count only.1 Reject Handling with the Sequential File Stage The Sequential File stage can optionally include a reject link. No part of this publication may be reproduced. 3. These actions include: 1. Push back – rows are sent to a Data Steward for corrective action. The default action is to push back reject and error rows to a Data Steward. stored in a retrieval system. transcribed.rej file extension. 3. and should be reviewed by the Data Steward. A message is always written to the Director log which details the count of rows successfully read and rows rejected. 2. Alternatively. By default. 2006 44 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. (Reject link exists) The reject option should be used in all cases where active management of the rejects is required. The Sequential File stage offers the following reject options: Option Description Continue Drop read failures from input stream. Pass successful reads to the output stream. Send the reject stream to a *. (No reject link exists) Fail Abort job on read format failure (No reject link exists) Output Reject switch failures to the reject stream. Use the Fail option. Use the Output option. .

err file or tag and merge with the output stream.rej or *. If there are multiple validations to perform. Rejects should not exist and should stop the job. Rejects should not exist but should not stop the job. 2006 45 of 179 © 2006 IBM Information Integration Solutions. . a local error handler based on a shared container can be used. they only need to be recorded but not acted on. and rejects can occur if the key fields are not found in the reference data. Use the Fail option. No part of this publication may be reproduced. Pass successful lookups to the output stream. Send the reject stream to an *. DS/EE offers the following options within a Lookup stage: Option Description Continue Ignore lookup failures and pass lookup fields as nulls to the output stream. This behavior makes the Lookup stage very valuable for positive (reference is found) and negative (reference is NOT found) business rule validation.rej file or tag and merge with the output stream.3 Reject Handling with the Transformer Stage Rejects occur when a transformer stage is used and a row: Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. stored in a retrieval system. transmitted. it must have a *. or translated into any language in any form by any means without the written permission of IBM. Alternatively. If a file is created by this option. the *.7.Information Integration Solutions Center of Excellence 3. Pass successful lookups to the output stream. Drop Drop lookup failures from the input stream. The reject option should be used in all cases where active management of the rejects is required. each must be done in its own Lookup.err file extension. and should be reviewed by the Data Steward. Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Rejects can exist in the data. All rights reserved.rej extension is used when rejects require investigation after a job run.7. Fail Abort job on lookup failure Reject Reject lookup failures to the reject stream. Furthermore. Pass successful lookups to the output stream. however. Lookup Stage Option Drop if lookup fields are necessary down stream or Continue if lookup fields are optional Send the reject stream to an *.err extension when rejects can be ignored but need to be recorded. to enforce error management ONLY ONE REFERENCE LINK is allowed on a Lookup stage. The *.2 Reject Handling with the Lookup Stage The Lookup stage compares a single input stream to one or more reference streams using keys. 2 3 4 3.

they only need to be recorded but not acted on.err extension when rejects can be ignored but need to be recorded. a shared container error handler can be used. The *.rej file or tag and merge with the output stream. OR 2. and Oracle Enterprise) offer an optional reject link that can be used to capture rows that cannot be written to the target database. however. To capture rejects from a target database. and be reviewed by the Data Steward. stored in a retrieval system. Rejects should not exist and should stop the job. . Rejects are categorized in the ETL job design document using the following ranking: Category 1 2 3 4 Description Rejects are expected and can be ignored. The reject option should be used in all cases where active management of the rejects is required.Information Integration Solutions Center of Excellence 1. Alternatively. If a file is created from the reject stream. Cannot satisfy requirements of any conditional output stream and is rejected by the default output stream. Rejects can exist in the data. a reject link must exist on that stage. Target database stages offer the following reject options: Option No reject link exists Reject link exists Description Do not capture rows that fail to be written. Transformer Stage Option Funnel the reject stream back to the output stream(s). or translated into any language in any form by any means without the written permission of IBM. the *. A message is always written to the Director log which details the count of rows successfully read and rows rejected. Alternatively.rej extension is used when rejects require investigation after a job run. Parallel Framework Red Book: Data Flow Job Design July 17. Send the reject stream to an *.4 Reject Handling with target database stages Some database stages (such as DB2/UDB Enterprise. it must have a *. ODBC Enterprise.err file extension.7. No part of this publication may be reproduced. it must have a *. All rights reserved. Send the reject stream to an *. transmitted. If a file is created by this option. Send the reject stream to a reject file and halt the job.rej or *.err file or tag and merge with the output stream.rej file extension. 3. a shared container error handler is used. Otherwise. 2006 46 of 179 © 2006 IBM Information Integration Solutions. Rejects should not exist but should not stop the job. transcribed. reject rows will not be captured. Satisfies requirements for a reject conditional output stream. Pass rows that fail to be written to the reject stream.

or translated into any language in any form by any means without the written permission of IBM.”. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The appropriate file type.rej file.5. The project phase (jpENVIRON) and a underscore “_”. The job name (jpJOB_NAME) and a underscore “_”. one of “rej” or “err”.rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504. stored in a retrieval system. This section deals with both methods of handling errors. and should be reviewed by the Data Steward. For example. 3. 3. 2.7. Reject link exists. and 5. The Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job. The project name (jpPROJECT_NAME) and a underscore “_”. job DECRP_N_XformClients in the ECR_FACTS project in the development environment with a serial number of 20060201-ETL-091504 would have these reject and error file names: ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504.err Rows will be converted to the common file record format with 9 columns (below) using Column Export and Transformer stages for each reject port.1 Processing Errors and Rejects to a Flat File Each job will produce a flat file for errors and a flat file for rejects with a specific naming convention: 1.5 Error Processing Requirements Jobs will produce flat files containing reject and errors and may alternatively process rows on reject ports and merge these rows with the normal output stream. The standard columns for error and reject processing are: Column Name HOST_NAME Key? Yes Data Source DSHostName transformer macro in the error handler July 17. All rights reserved. The job serial number (jpJOBSERIALNO) and a period “.7. transcribed.Rejects are tracked by count only. . 2 Rejects should not exist but should not stop the job. Only records that match the given table definition and database constraints are written.Information Integration Solutions Center of Excellence Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Target Database Stage Option No reject link exists. and gathered using a Funnel stage that feeds a Sequential File stage. 4. 3. transmitted. Send the reject stream to a *. No part of this publication may be reproduced.

The Transformer stage adds the required key columns. 2006 48 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. . transmitted. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence PROJECT_NAME JOB_NAME STAGE_NAME DATA_OBJ_NAME JOB_SERIALNO ETL_ROW_NUM ETL_BAT_ID ROW_DATA Yes Yes Yes Yes Yes Yes Yes No DSProjectName transformer macro in the error handler DSJobName transformer macro in the error handler The name of the stage from which the error came The source table or file data object name jpJOBSERIALNO Data stream coming in to the error handler Data stream coming in to the error handler The columns from the upstream stages reject port exported to a single pipe-delimited “|” varchar(2000) column using the Column Export stage in the error handler In this example. transcribed. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. the following stages process the only errors produced by a job: The Column Export stage maps the unique columns to the single standard column. Figure 13: Error Processing Components Parallel Framework Red Book: Data Flow Job Design July 17.

ROW_DATA: Figure 14: Error Processing Column Export stage And the downstream Transformer stage builds the standard output record by creating the required keys: Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. or translated into any language in any form by any means without the written permission of IBM. 2006 49 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence The input to the Column Export stage explicitly converts the data unique to the reject stream (in this case. All rights reserved. Track*) to a single output column. transmitted. . stored in a retrieval system.

Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3.7. Connect the reject port to a Transformer stage where columns are set to specific values. Connect the output stream of the Transformer stage and one or more output streams of the Switch stage to a Funnel stage to merge the two (or more) streams. stored in a retrieval system. All rights reserved. The following table identifies the tagging method to be used for the previously cited operators. transmitted. 2006 50 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. OR with columns contain illegal values for some operation performed on said columns. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. No part of this publication may be reproduced. This is done by processing the rows from the reject ports and setting the value of a specific column with a value specified by the design document. rows rejected by the Lookup stage are processed by a corrective Transformer stage where the failed references as set to a specific value and then merged with the output of the Lookup stage: Parallel Framework Red Book: Data Flow Job Design July 17. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream. Switch Transformer In this example. Connect the output stream of the corrective Transformer stage and one or more output streams of the original Transformer stage to a Funnel stage to merge the two (or more) streams. A failed switch will reject an intact input row show key fails to resolve to one of the Switch output stream. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key. transcribed. . Connect the output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams. In either case.2 Processing Errors and Rejects and Merging with an Output Stream There may be processing requirements that specify that rejected or error rows be tagged as having failed a validation and merged back into the output stream. One or more columns may have been selected for replacement when a reference key is found. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values.5. Connect the reject port to a Transformer stage where columns are set to specific values.

All rights reserved. stored in a retrieval system. transmitted. or translated into any language in any form by any means without the written permission of IBM. This section provides guidelines appropriate use of various stages when building a parallel job flows. transcribed. 3.Information Integration Solutions Center of Excellence Figure 16: Error Processing Lookup example 3. 2006 51 of 179 © 2006 IBM Information Integration Solutions. In clustered an MPP environments Server Edition components only run on the primary (conductor) node.8 Component Usage DataStage Enterprise Edition offers a wealth of component types for building ETL flows. No part of this publication may be reproduced. The ability to use a Server Edition component within a parallel job is intended only as a migration option for existing Server Edition applications that might benefit by leveraging some parallel capabilities on SMP platforms.8. Server Edition components limit overall performance of large-volume job flows since many components such as the BASIC Transformer use interpreted psuedo-code. Parallel Framework Red Book: Data Flow Job Design July 17. severely impacting scalability and network resources. .1 Server Edition Components Avoid the use of Server Edition components in parallel job flows.

3. • The Copy stage should be used instead of a Transformer for simple operations including: .BASIC Transformers .8.8. o Unless the Force property is set to “True”. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run. Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form. and necessary. it is best to develop a job iteratively using the Copy stage as a “placeholder”. in sort order. it can be used at the end of a data flow o For simple jobs with only two stages. it is important to minimize the number of transformers. always write to parallel Data Sets.2 Copy Stage For complex data flows.Job Design placeholder between stages (unless the Force option =true. 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Data Sets offer parallel I/O on read and write operations.Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: . Management. and in Enterprise Edition native format.4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. the Copy stage should be used as a placeholder so that new stages can be inserted easily should future requirements change. No part of this publication may be reproduced. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime. NOTE: Because parallel Data Sets are platform and configuration-specific.Renaming Columns . 3. Enterprise Edition will optimize this out at runtime) . and to use other stages (such as Copy) when derivations are not needed.8. or translated into any language in any form by any means without the written permission of IBM. and Production Automation. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs.3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs. This is discussed in more detail in Parallel Framework Standard Practices: Administration.BASIC Routines . transmitted. All rights reserved. transcribed. Since the Copy stage does not require an output link. 3. Used in this manner. For this reason. without overhead for format or data type conversions. . stored in a retrieval system. they should not be used for long-term archive of source data.Dropping Columns July 17.Server shared containers Note that BASIC Routines are still appropriate.

5 BuildOp stages BuildOps should only be used when: . and character string trimming.6. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible. No part of this publication may be reproduced. Consider. • NEVER use the “BASIC Transformer” stage in large-volume job flows. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines. 2006 53 of 179 © 2006 IBM Information Integration Solutions. See Section 8. The only time that Filter or Switch should be used is when the selection clauses need to be parameterized at runtime.2: Default and Explicit Type Conversions] Note that rename. and default type conversion can also be performed by the output mapping tab of any stage. 3.5. • Because the parallel Transformer is compiled. transmitted. implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. or translated into any language in any form by any means without the written permission of IBM. performance should this should be tested in isolation to identify specific cause of bottlenecks.2: Modify Stage.3 Then B=”X” If A=4.2.8. For example. null handling. Even then its use should be restricted and the routines should be converted as soon as possible.1. . drop (if Runtime Column Propagation is disabled). transcribed. it is faster than the interpreted Filter and Switch stages. stored in a retrieval system. if possible.1. Parallel Framework Red Book: Data Flow Job Design July 17. Instead.Existing Transformers do not meet performance requirements As always. user-defined functions and routines can expand parallel Transformer capabilities. • The Modify stage can be used for non-default type conversions.Complex reusable logic cannot be implemented using the Transformer or .7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B. the derivation expression: If A=0. All rights reserved.Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4.

Numeric Float. unit32 int64. unit64 raw 4 bytes (roundup(p)+1)/2 4 bytes 8 bytes 1 byte 2 bytes 4 bytes 8 bytes 1 byte per character Date with month.Information Integration Solutions Center of Excellence 4 DataStage Data Types The DataStage Designer and Manager represent column data types using SQL notation. LongVarChar. or translated into any language in any form by any means without the written permission of IBM. IEEE single-precision (32-bit) floating point value IEEE double-precision (64-bit) floating point value Signed or unsigned integer of 8 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 16 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 32 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 64 bits (Specify unsigned Extended option for unsigned) Untyped collection. compatible with IBM packed decimal format. with resolution of microseconds (Specify microseconds Extended option) Single field containing both date and time value with resolution to seconds. Bit. stored in a retrieval system. LongVarChar. Single field containing both date and time value with resolution to microseconds. VarChar Time Time Timestamp Timestamp date decimal sfloat dfloat int8. 2006 54 of 179 © 2006 IBM Information Integration Solutions. The following table summarizes the underlying data types of DataStage Enterprise Edition: SQL Type Internal Type Size Description Date Decimal. The internal Enterprise Edition data types are used in schema files and are displayed when viewing generated OSH or viewing the output from $OSH_PRINT_SCHEMAS. VarChar NChar. transcribed. LongVarBinary. LongNVarChar Char. Parallel Framework Red Book: Data Flow Job Design July 17. uint16 int32. and year Packed decimal. consisting of a fixed or variable number of contiguous bytes and an optional alignment value ASCII character string of fixed or variable length (Unicode Extended option NOT selected) ASCII character string of fixed or variable length (Unicode Extended option NOT selected) ASCII character string of fixed or variable length (Unicode Extended option IS selected) Time of day. transmitted. NVarChar. Char. VarBinary Unknown. No part of this publication may be reproduced. day. Real Double TinyInt SmallInt Integer BigInt1 Binary. (Specify microseconds Extended option) string 1 byte per character ustring ustring time time(microseconds) timestamp timestamp(microseconds) multiple bytes per character multiple bytes per character 5 bytes 5 bytes 9 bytes 9 bytes 1 BigInt values map to long long integers on all supported platforms except Tru64 where they map to longer integer values. . All rights reserved. Each SQL data type maps to an underlying data type in the Enterprise Edition engine. with resolution to seconds Time of day. uint8 int16.

2 Default and Explicit Type Conversions DataStage Enterprise Edition provides a number of default conversions and conversion functions when mapping from a source to a target data type.) An Extended field appears in the columns grid. VarChar. A blank cell indicates that no conversion is provided. String data represents unmapped bytes. specify that these data types are extended. date uint8 sfloat uint16 uint32 uint64 dfloat int16 int32 int64 string ustring time e e e int8 timestamp e e e e 55 of 179 decimal raw Int8 uint8 Int16 uint16 Int32 uint32 Int64 uint64 sfloat dfloat decimal string d de d de d de d de de de de d d d d d d d d d d d d d d d d d d d d d de d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d de d d d d d d d d de de d d d d d d d d d d e d d d d d d d d d d d d d d de d d d d d d de d d d de d d d de d d e d d e d e d e d e d d d d e d e de d de de de de d d d de de d e e e e Parallel Framework Red Book: Data Flow Job Design July 17. You can.1.Information Integration Solutions Center of Excellence 4. stored in a retrieval system. 4. NVarChar. 2006 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. e = You can use a Modify or a Transformer conversion function to explicitly convert from the source field type to the destination field type. parallel jobs support two types of underlying character data types: strings and ustrings. The NChar. and LongVarChar SQL types relate to underlying string types where each character is 8-bits and does not require mapping because it represents an ASCII character.1 Strings and Ustrings If NLS is enabled on your DataStage server. in which case they are taken as ustrings and do require mapping. . ustring data represents full Unicode (UTF-16) data. transcribed. transmitted.1. The Char. The following table summarizes Data Type conversions: Source Field Target Field d = There is a default type conversion from source field type to destination field type. however. VarChar. or LongVarChar columns have ‘Unicode’ in this field. and LongNVarChar types relate to underlying ustring types so do not need to be explicitly extended. and extended Char. (They are specified as such by selecting the Extended check box for the column in the Edit Meta Data dialog box. No part of this publication may be reproduced. Default type conversions take place across the stage output mappings of any Enterprise Edition stage.

for example. . No part of this publication may be reproduced.2 Null Handling DataStage Enterprise Edition represents nulls in two ways: . When used in these stages. When converting from variable-length to fixed-length strings. Enterprise Edition pads the remaining length with NULL (ASCII zero) characters by default. You must first convert a Char string type to a Varchar type before using PadString. When entering a space for the value of APT_STRING_PADCHAR do note enclose the space character in quotes. In-band null representation can be disadvantageous because you must reserve a field value for nulls and this value cannot be treated as valid data elsewhere. Some stages (for example. stored in a retrieval system.It allocates a single bit to mark a field as null. This type of representation is called an in-band null. In these instances. the specified pad character will override the default for that stage only. This type of representation is called an out-of-band null. • The environment variable APT_STRING_PADCHAR can be used to change the default pad character from an ASCII NULL (0x0) to another character. Enterprise Edition displays a warning message in the job log. All rights reserved. the PadString Transformer function can be used to pad a variablelength (Varchar) string to a specified length using a specified pad character. or translated into any language in any form by any means without the written permission of IBM. transcribed. The Transformer and Modify stages can change a null representation from an out-of-band null to an inband null and from an in-band null to an out-of-band null.It designates a specific field value to indicate a null.Information Integration Solutions Center of Excellence ustring raw date time timesta mp de e e e e d de e d d e e e e de e d d d de e d e d e e e e e e e e e de e e e e The conversion of numeric data types may result in a loss of precision and cause incorrect results. an ASCII space (0x20) or a Unicode space (U+0020). for example a numeric field’s most negative possible value. • 4. Parallel Framework Red Book: Data Flow Job Design July 17. • As an alternate solution. transmitted. . 2006 56 of 179 © 2006 IBM Information Integration Solutions. depending on the source and result data types. Note that PadString does not work with fixed-length (CHAR) string types. Sequential File and DB2/UDB Enterprise targets) allow the pad character to be specified in their stage or column definition properties.

destination value is never null. When reading from Data Set and database sources with nullable columns. See Section 8. transmitted. July 17. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. Source Field not Nullable Nullable not Nullable Nullable Destination Field not Nullable Nullable Nullable not Nullable Result Source value propagates to destination. .1. care must be taken to avoid data rejects. the source value propagates. All rights reserved. stored in a retrieval system. as shown in Figure 17: Figure 17: Extended Column Metadata (Nullable properties) The Table Definition of a stage’s input or output data set can contain columns defined to support outof-band nulls (Nullable attribute is checked).Information Integration Solutions Center of Excellence IMPORTANT: When processing nullable columns in a Transformer stage. Enterprise Edition uses the internal.1: Transformer NULL Handling and Reject Link. Source value or null propagates. transcribed. No part of this publication may be reproduced. The next table lists the rules for handling nullable fields when a stage takes a Data Set as input or writes to a Data Set as output. 2006 57 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. out-of-band null representation for NULL values. Source value propagates. If the source value is not null. When reading from or writing to Sequential Files or File Sets.

Information Integration Solutions Center of Excellence If the source value is null. Some stages. for example the Sequential File stage. All rights reserved. only the columns explicitly referenced within the shared container logic need to be defined. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 58 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. allow their runtime schema to be parameterized further extending re-use through RCP. Using RCP judiciously in a job design facilitates re-usable job designs based on input metadata. Using RCP. No part of this publication may be reproduced. RCP facilitates re-use through parallel shared containers. the remaining columns pass through at runtime. it must be enabled at the project level through the Administrator client. Furthermore. 4. . transmitted.3 Runtime Column Propagation Runtime column propagation (“RCP”) allows job designs to accommodate additional columns beyond those defined by the job developer. as long as each stage in the shared container has RCP enabled on their stage Output properties. stored in a retrieval system. a fatal error occurs. rather than using a large number of jobs with hard-coded table definitions to perform the same tasks. Before a DataStage developer can use RCP.

transcribed. Keyed partitioning examines the data values in one or more key columns. Remove Duplicates) or stage requirements (for example.1 Partition Types While partitioning allows data to be distributed across multiple processes running in parallel. Stage running sequentially Stage running in parallel 5. ensuring that records with the same values in those key column(s) are assigned to the same partition. Different types of keyless partitioning methods define the method of data distribution. and are indicated by a “fan-in” icon as shown in this example: Stage running in parallel Figure 19: Collector icon This section provides an overview of partitioning and collecting methods. It also provides tips for monitoring jobs running in parallel. 2006 59 of 179 © 2006 IBM Information Integration Solutions. Join) require processing on groups of related records. Parallel Framework Red Book: Data Flow Job Design July 17. different types of partitioning are provided for the parallel job developer.Keyless partitioning distributes rows without regard to the actual data values. transmitted. . a “fan-out” icon is drawn on the link within the Designer canvas. If the prior stage was running sequentially. Partitioners exist before any stage that is running in parallel. Collectors only exist before stages running sequentially and when the previous stage is running in parallel. No part of this publication may be reproduced. stored in a retrieval system. For this reason. Keyed partitioning is used when business rules (for example.Information Integration Solutions Center of Excellence 5 Partitioning and Collecting Partition parallelism is a key to establishing scalable performance of DataStage Enterprise Edition. or translated into any language in any form by any means without the written permission of IBM. and provides guidelines for appropriate use in job designs. Partitioners distribute rows of a single link into smaller segments that can be processed independently in parallel. it is important that this distribution does not violate business requirements for accurate data processing. as shown in this example: Stage running sequentially Figure 18: “fan-out” icon Collectors combine parallel partitions of a single link for sequential processing. All rights reserved. Partitioning methods are separated into keyless and keyed classes: .

The “Preserve Partitioning” flag is an internal “hint” that Auto partitioning uses to attempt to preserve carefully ordered data (for example. then a keyed partitioning method must be specified to achieve correct results. Auto partitioning will ensure correct results when using built-in stages. transcribed.1 Auto Partitioning The default partitioning method for newly-drawn links. .Information Integration Solutions Center of Excellence The default partitioning method for newly-drawn links is Auto partitioning. However. on the output of a parallel Sort). Based on the configuration file.1. the partitioning method may not necessarily be the most efficient from an overall job perspective. to improve performance. Entire) and keyed (Hash) partitioning methods to produce functionally correct results and. Within the Designer canvas. All rights reserved. transmitted. Auto partitioning will select between keyless (Same. However. The partitioning method is specified in the Input stage properties using the “Partitioning” option as shown on the right: Figure 20: Specifying Partition method 5. or translated into any language in any form by any means without the written permission of IBM. links with Auto partitioning are drawn with the following link icon: Figure 21: Auto partitioning icon Auto partitioning is designed to allow the beginning DataStage developer to construct simple data flows without having to understand the details of parallel design principles. 2006 60 of 179 © 2006 IBM Information Integration Solutions. This flag is set automatically by Parallel Framework Red Book: Data Flow Job Design July 17. the ability for the Enterprise Edition engine to determine the appropriate partitioning method depends on the information available to it. For example. No part of this publication may be reproduced. Round Robin. Auto partitioning specifies that the Enterprise Edition engine will attempt to select the appropriate partitioning method at runtime. Data Sets. stored in a retrieval system. in some cases. since the Enterprise Edition engine has no visibility into user-specified logic (such as Transformer or BuildOp stages) it may be necessary to explicitly specify a partitioning method for some stages. Furthermore. In general. and job design (stage requirements and properties). if the logic defined in a Transformer stage is based on a group of related records.

Because Same does not redistribute existing partitions. stored in a retrieval system. In these instances. in a round robin partition assignment. .2.1.2 Keyless Partitioning Keyless partitioning methods distribute rows without examining the contents of the data: Keyless Partition Method Same Round Robin Random Entire Description Retains existing partitioning from previous stage. or translated into any language in any form by any means without the written permission of IBM. a warning will be placed in the Director log indicating that Enterprise Edition was unable to preserve partitioning for a specified stage. 2006 61 of 179 © 2006 IBM Information Integration Solutions. transcribed. Distributes rows evenly across partitions. and its state is stored in persistent Data Sets. Each partition receives the entire Data Set. as shown in on the right: Figure 22: Preserve Partitioning option The Preserve Partitioning flag is part of the Data Set structure. if the Preserve Partitioning flag was set. 5. although it can be explicitly set or cleared in the “Advanced” stage properties of a given stage. a parallel Sort).1 Same Partitioning Same partitioning in fact performs no partitioning to the input Data Set. it retains the partitioning from the output of the upstream stage. for example). There are some cases when the input stage requirements prevent partitioning from being preserved. Row ID's 0 3 6 1 4 7 2 5 8 0 3 6 1 4 7 2 5 8 Within the Designer canvas. and is appropriate when trying to preserve the grouping of a previous operation (for example. Distributes rows evenly across partitions in a random partition assignment. as illustrated on the right: Same partitioning doesn’t move data between partitions (or.1. in the case of a cluster or Grid. 5. links that have been specified with Same partitioning are drawn with a “horizontal line” partitioning icon: Figure 23: Same partitioning icon It is important to understand the impact of Same partitioning in a given data flow. the degree of parallelism remains unchanged: Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. No part of this publication may be reproduced. Instead.Information Integration Solutions Center of Excellence some stages (Sort. between servers). All rights reserved.

as illustrated on right: Entire partitioning is useful for distributing the reference data of a Lookup task (this may or may not involve the Lookup stage). or translated into any language in any form by any means without the written permission of IBM.2. transmitted. regardless of the current $APT_CONFIG_FILE 5. as the complete Data Set must be distributed across the network to each node. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). . 3 2 1 0 . the downstream stage runs with the degree of parallelism used to create the Data Set. . Entire partitioning may have a performance impact. 2006 62 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5. No part of this publication may be reproduced.1. …8 7 6 5 4 3 2 1 0 ENTIRE . Random partitioning has a slightly higher overhead than Round Robin partitioning. Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning. but using a random assignment.2. the order that rows are assigned to a particular partition will differ between job runs. . While in theory Random partitioning is not subject to regular data patterns that might exist in the source data. similar to dealing cards: Round robin partitioning has a fairly low overhead.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially.2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment. 5. stored in a retrieval system. Since the random partition number must be calculated. Since optimal parallel processing occurs when all partitions have the same workload. As a result. On clustered and Grid implementations.1. transcribed. 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17. 3 2 1 0 .1. .2.3 Random Partitioning Like Round Robin. Random partitioning evenly distributes rows across partitions. it is rarely used in real-world data flows.

5.1.3. 2006 63 of 179 © 2006 IBM Information Integration Solutions. Join) require processing on groups of related records. and there are a large number of unique values. Assigns rows with the same values in one or more key column(s) to the same partition using a specified range map generated by pre-reading the Data Set. Remove Duplicates) or stage requirements (for example. consider the following sample Data Set: ID 1 2 3 4 5 6 7 8 9 10 LName Ford Ford Ford Ford Dodge Dodge Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Horace John Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 17840 Jefferson 75 Boston Boulevard 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Values of key column …0 3 2 1 0 2 3 2 1 1 HASH 0 3 0 3 1 1 1 2 2 2 Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. . transmitted. ensuring that records with the same values in those key column(s) are assigned to the same partition. All rights reserved.1 Hash Partitioning Hash partitioning assigns rows with the same values in one or more key column(s) to the same partition using an internal hashing algorithm.Information Integration Solutions Center of Excellence 5. For DB2 Enterprise Server Edition with DPF (DB2/UDB) only – matches the internal partitioning of the specified source or target table. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. then the resulting partitions will be of relatively equal size. If the source data values are evenly distributed within these key column(s). As an example of hashing. Keyed partitioning is used when business rules (for example. Keyed Partition Method Hash Modulus Range DB2 Description Assigns rows with the same values in one or more key column(s) to the same partition using an internal hashing algorithm.1. stored in a retrieval system. transcribed.3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns. No part of this publication may be reproduced.

All rights reserved.Information Integration Solutions Center of Excellence Hashing on key column LName would produce the following results: Partition 0: ID 5 6 LName Dodge Dodge FName Horace John Address 17840 Jefferson 75 Boston Boulevard Partition 1: ID 1 2 3 4 7 8 9 10 LName Ford Ford Ford Ford Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. 2006 64 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. . transmitted. No part of this publication may be reproduced. stored in a retrieval system.

Using the same source Data Set. or translated into any language in any form by any means without the written permission of IBM. producing partition skew. hash partitioning on the key columns LName and FName yields the following distribution with a 4-node configuration file: Partition 0: I D 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen I D 4 6 10 LName Ford Dodge Ford Partition 2: FName Eleanor John Eleanor Address 7900 Jefferson 75 Boston Boulevard 1100 Lakeshore Partition 1: I D 3 5 9 LName Ford Dodge Ford FName Edsel Horace Edsel Address 7900 Jefferson 17840 Jefferson 1100 Lakeshore I D 1 7 LName Ford Ford Partition 3: FName Henry Henry Address 66 Edison Avenue 4901 Evergreen In this example. transcribed. there are more instances of “Ford” than “Dodge”. transmitted. individual key column values have no significance for partition assignment. Also note that only the unique combination of key column values appear in the same partition when used for hash partitioning. 2006 65 of 179 © 2006 IBM Information Integration Solutions. the partition size of modulus partitioning will be equally distributed as long as the data values in the key column are equally distributed.3. The remainder is used to assign the value to a given partition: partition = MOD (key_value / number of partitions) Like hash. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. . 5.1. regardless of the actual number of nodes in the parallel configuration file. which would impact performance. When using hash partitioning on a composite key (more than one key column). It performs a modulus operation on the data value using the number of partitions as the divisor. Also note that in this example the number of unique values will limit the degree of parallelism. All rights reserved.Information Integration Solutions Center of Excellence In this case.2 Modulus Partitioning Modulus partitioning uses a simplified algorithm for assigning related records based on a single integer key column. No part of this publication may be reproduced. the key column combination of LName and FName yields improved data distribution and a greater degree of parallelism.

No part of this publication may be reproduced. To maintain partitioning on data read from a DB2/UDB Enterprise stage. data is read in parallel from each DB2 node. 5. It is important to note that if the data distribution changes without recreating the Range Map. depending on the value. Range partitioning must read the Data Set twice: once to create a Range Map file.1.4 DB2 Partitioning The DB2/UDB Enterprise stage matches the internal database partitioning of the source or target DB2 Enterprise Server Edition with Data Partitioning Facility database (previously called “DB2/UDB EEE”). Values of key column 4 0 5 1 6 0 5 4 3 RANGE Rang e Map fi le 0 1 0 4 4 3 The “read twice” penalty of Range partitioning limits its use to specific scenarios. if the incoming Data Set is sequential and ordered on the key column(s). All rights reserved. it should be used if you have a single integer key column. Also. transcribed. 2006 66 of 179 © 2006 IBM Information Integration Solutions. or for a non-integer key column.3 Range Partitioning As a keyed partitioning method. unlike Hash and Modulus partitioning where partition skew is dependent on the actual data distribution. and the second to actually partition the data within a flow using the Range Map. To achieve this balanced distribution. these rows will be assigned to either the first or the last partition. Modulus partitioning cannot be used for composite keys. typically where the incoming data values and distribution are consistent over time. if new data values are processed outside of the range of a given Range Map. defeating the intention of Range partitioning. by default. . In these instances.3. Given a sufficient number of unique values.Information Integration Solutions Center of Excellence Since modulus partitioning is simpler and faster than hash. 5. Using the DB2/UDB Enterprise stage.1. partition balance will be skewed. the Range Map file can be re-used. Range partitioning will result in sequential processing.3. DB2 partitioning can only be specified for target DB2/UDB Enterprise stages. use Same partitioning on the input to downstream stages. A Range Map file is specific to a given parallel configuration file. transmitted. And. Range partitioning ensures balanced workload by assigning an approximately equal number of rows to each partition. data is partitioned to match the internal partitioning of the target DB2 table using the DB2 partitioning method. Range partitioning assigns rows with the same values in one or more key column(s) to the same partition. when writing data to a target DB2 database using the DB2/UDB Enterprise stage. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. In another scenario to avoid. Parallel Framework Red Book: Data Flow Job Design July 17.

starting at zero. transmitted. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system.Information Integration Solutions Center of Excellence 5.2Understanding the Parallel Job Score. and select the “Show Instances” option. 2006 67 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. . right-click anywhere in the window. as illustrated below. if specified. in some cases. Specific details on interpreting the parallel job score can be found in 12. and each stage is processing an equal number (12.2 Monitoring Partitions At runtime. Figure 25: Director Job Monitor row counts by partition Setting the environment variable APT_RECORD_COUNTS will output the row count per link per partition to the Director log as each stage/node completes processing. Partitions are assigned numbers. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name). transcribed. or translated into any language in any form by any means without the written permission of IBM.4. The partition number is appended to the stage name for messages written to the Director log. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. This is very useful in determining the distribution across parallel partitions (skew). In this instance. databases) c) and. No part of this publication may be reproduced. DataStage Enterprise Edition determines the degree of parallelism for each stage using: a) the parallel configuration file (APT_CONFIG_FILE) b) the degree of parallelism of existing source and target Data Sets (and. as shown in the example log below where the stage named “Peek” is running with four degrees of parallelism (partition numbers zero through 3): Figure 24: Partition numbers as shown in Director log To display row counts per partition in the Director Job Monitor window.500) of rows for an optimal balanced workload.

transmitted. the following objectives help to form a methodology for assigning partitioning: Objective 1: Choose a partitioning method that gives close to an equal number of rows in each partition. 5.3 Partition Methodology Given the numerous options for keyless and keyed partitioning. or Manager) can be used to identify the degree of parallelism and number of rows per partition for an existing persistent Data Set. This ensures that the processing workload is evenly balanced.Information Integration Solutions Center of Excellence Figure 26: Output of APT_RECORD_COUNTS in Director log Finally. the “Data Set Management” tool (available in the Tools menu of Designer. as shown below: Figure 27: Data Set Management Tool In a non-graphical way. the orchadmin command line utility on the DataStage server can also be used to examine a given parallel Data Set. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 68 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. while minimizing overhead. . minimizing overall run time. transcribed. stored in a retrieval system. No part of this publication may be reproduced. Director.

Change Apply. Objective 4: Partition method should not be overly complex The simplest method that meets the above objectives will generally be the most efficient and yield the best performance. Remove Duplicates. Using the above objectives as a guide. but is not limited to: Aggregator. Note that in satisfying the requirements of this second objective.Information Integration Solutions Center of Excellence Objective 2: The partition method must match the business requirements and stage functional requirements. No part of this publication may be reproduced. transcribed. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. Objective 3: Unless partition distribution is highly skewed. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. and Sort stages. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). persistent Data Sets can be used to retain the partitioning and sort order. . 2006 69 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. Join. Merge. o Across jobs. use Round Robin partitioning to redistribute data equally across all partitions o Especially useful if the input Data Set is highly skewed or sequential d) Use Same partitioning to optimize end-to-end partitioning and to minimize repartitioning o Being mindful that Same partitioning retains the degree of parallelism of the upstream stage o Within a flow. Change Capture. transmitted. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. or translated into any language in any form by any means without the written permission of IBM. assigning related records to the same partition if required Any stage that processes groups of related records (generally using one or more key columns) must be partitioned using a keyed partition method. minimize repartitioning. stored in a retrieval system. This includes. It may also be necessary for Transformers and BuildOps that process groups of related records. This is particularly useful if downstream jobs are run with the same degree of parallelism (configuration file) and require the same partition and sort order. the following methodology can be applied: a) Start with Auto partitioning (the default) b) Specify Hash partitioning for stages that require groups of related records o Specify only the key column(s) that are necessary for correct grouping as long as the number of unique values is sufficient o Use Modulus partitioning if the grouping is on a single integer key column o Use Range partitioning if the data is highly skewed and the key column values and distribution do not change significantly over time (Range Map can be reused) c) If grouping is not required.

2 Partitioning Example 2 – Use of Entire Partitioning In this example. The output of the Aggregator is then sent to the second input of the Join.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. or translated into any language in any form by any means without the written permission of IBM. 5.4. No part of this publication may be reproduced. or through Auto partitioning): Parallel Framework Red Book: Data Flow Job Design July 17. a Transformer is used to extract data from a single header row of an input file. To add aggregate columns to every detail row. we can move the Hash partition and Sort before the Copy stage. 5. Within the Transformer. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row.4 Partitioning Examples In this section. the partitioning and sorting of this scenario can be optimized. stored in a retrieval system. and apply Same partitioning to the downstream links. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator. . Using a “standard” solution. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly. more advanced partitioning and sorting examples are given in 12. transcribed. The standard solution would be to Hash partition (and Sort) the inputs to the Join and Aggregator stages as shown below: Figure 28: “Standard” Partitioning assignment However. as shown below: Figure 29: Optimized Partitioning assignment This example will be revisited in the Sorting discussion because there is one final step necessary to optimize the sorting in this example. a new output column is defined on the header and detail links using a single constant value derivation. we’ll apply the partitioning methodology defined earlier to several example job flows. All rights reserved. transmitted. 2006 70 of 179 © 2006 IBM Information Integration Solutions.2Sorting and Hashing Advanced Example.4. on closer inspection. Additional.Information Integration Solutions Center of Excellence 5.4. Because the Join and Aggregator use the same partition keys and sort order.

The Join stage operates by reading a single row from the Left input and reading all rows from the Right input that match the key value(s). stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. the above solution has one serious limitation. the single value join column will assign all rows to a single partition. . transmitted. transcribed. depending on the business requirements. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 30: “Standard” Partitioning assignment for a Join stage Although Hash partitioning guarantees correct results for stages that require groupings of related records. the link order in this example should be set so that the single header row is assigned to the Right input. and the detail rows are assigned to the Left input as shown in the following illustration: Figure 32: Specifying Link Order in Join stage Parallel Framework Red Book: Data Flow Job Design July 17.Use Entire partitioning on the header input to copy the single header row to all partitions Figure 31: Optimized Partitioning assignment based on business requirements Because we are joining on a single value. For this reason. so we will revisit this in the Sorting discussion. the link order of the Inner Join is significant. consider that the single header row is really a form of reference data. resulting in sequential processing. All rights reserved. In order to process a large number of detail records. it is not always the most efficient solution. there is no need to pre-sort the input to the Join. An optimized solution would be to alter the partitioning for the input links to the Join stage: . Although functionally correct. Remembering that the degree of parallel operation is limited by the number of distinct values. To optimize partitioning. 2006 71 of 179 © 2006 IBM Information Integration Solutions.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions .

Because the Join will wait until it receives an End of Group (new key value) or End of Data (no more rows on the input Data Set) from the Right input. All rights reserved. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition.5. the order of rows in an Auto collector is undefined. and may vary between job runs on the same Data Set. For this reason. (See Section 12.5. or translated into any language in any form by any means without the written permission of IBM. 5. Consider an example where data is read sequentially and passed to a Round Robin partitioner: Parallel Framework Red Book: Data Flow Job Design July 17. For advanced users. transmitted.2 Round Robin Collector The Round Robin collector patiently reads rows from partitions in the input Data Set by reading input partitions in round robin order.1 Auto Collector The Auto collector eagerly reads rows from partitions in the input Data Set without blocking if a row is unavailable on a particular partition. transcribed. the detail rows in the Left input will buffer to disk to prevent a deadlock. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence If defined in reverse of this order. the collector method is defined in the stage Input/Partitioning properties for any stage running sequentially. . 2006 72 of 179 © 2006 IBM Information Integration Solutions. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5.3: Minimizing Runtime Processes and Resource Requirements). Changing the output derivation on the header row to a series of numbers instead of a constant value will establish the End of Group and prevent buffering to disk. stored in a retrieval system. Like partitioning methods. Auto is the default collector method. there is a specialized example where the Round Robin collector may be appropriate.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially. there is one further detail in this example. 5. However.

the Sort Merge collector will generate a sequential stream of rows in globally sorted order. Row order is undefined for non-key columns.Information Integration Solutions Center of Excellence Round Robin partitioner Round Robin collector Sequential input Stage running in parallel Sequential output Figure 34: RoundRobin Collector example Assuming the data is not repartitioned within the job flow and that the number of rows is not reduced (for example. transcribed.3 Ordered Collector An Ordered collector reads all rows from the first partition.6 Collecting Methodology Given the options for collecting data into a sequential stream. In this scenario.5. then reads all rows from the next partition until all rows in the Data Set have been collected. through aggregation). then a Round Robin collector can be used before the final Sequential output to reconstruct a sequential output stream in the same order as the input data stream. as used to sort the input Data Set in parallel. stored in a retrieval system. use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. the Ordered collector may be more efficient c) Use a Round Robin collector to reconstruct rows in input order for round-robin partitioned input Data Sets. in the same order. or translated into any language in any form by any means without the written permission of IBM. 5. No part of this publication may be reproduced. use Sort Merge collector to produce a single.4 Sort Merge Collector If the input Data Set is sorted in parallel. Parallel Framework Red Book: Data Flow Job Design July 17. 5. All rights reserved. an Ordered collector will generate a sequential stream in sort order. 2006 73 of 179 © 2006 IBM Information Integration Solutions. transmitted. The Sort Merge collector requires one or more key column(s) to be defined. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter.5. Ordered collectors are generally only useful if the input Data Set has been Sorted and Range partitioned on the same key column(s). 5. This is because Round Robin collector reads from partitions using the same partition order that a Round Robin partitioner assigns rows to parallel partitions. as long as the Data Set has not been repartitioned or reduced. and these should be the same columns. . globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned.

This sort can be done in parallel. ordered result set. based on one or more key column(s) NOTE: By definition. In the following example. Change Apply.Sorting: is used to establish group order within each partition. the process of sorting data uses one primary key column and. 2006 74 of 179 © 2006 IBM Information Integration Solutions. optionally.1 Partition and Sort Keys Using the parallel Sort within DataStage Enterprise Edition: . Merge) require pre-sorted groups of related records. 6. one or more secondary key column(s) to generate a sequential. a sort is required after repartitioning. transmitted. Change Capture. sorting on primary key LName (ascending). The order of key columns determines the sequence and groupings in the result set. or translated into any language in any form by any means without the written permission of IBM. the Remove Duplicates stage selects either the first or last row from each group of an input Data Set sorted by one or more key columns. To restore row order and groupings. Instead. the previous input Data Set is partitioned on LName and FName columns. Given a 4-node configuration file. No part of this publication may be reproduced. Join. sort order is not maintained. For example. sorting is most often needed to establish order within specified groups of data. transcribed. This is the method the SQL databases use for an ORDER BY clause. assigning rows with the same key column values to the same partition . secondary key FName (descending): Input Data: I D 1 2 3 4 5 6 7 8 9 10 LName Ford Ford Ford Ford Dodge Dodge Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Horace John Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 17840 Jefferson 75 Boston Boulevard 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore I D 6 5 1 7 4 10 3 9 2 8 After Sorting by LName.Partitioning: is used to gather related records. . in most cases there is no need to globally sort data to produce a single sequence of rows.Information Integration Solutions Center of Excellence 6 Sorting Traditionally. Sort Aggregator. Each column is specified with an ascending or descending sort order. we would see the following results: Parallel Framework Red Book: Data Flow Job Design July 17. Other stages (for example. as illustrated in the following example. FName: LName Dodge Dodge Ford Ford Ford Ford Ford Ford Ford Ford FName John Horace Henry Henry Eleanor Eleanor Edsel Edsel Clara Clara Address 75 Boston Boulevard 17840 Jefferson 66 Edison Avenue 4901 Evergreen 7900 Jefferson 1100 Lakeshore 7900 Jefferson 1100 Lakeshore 66 Edison Avenue 4901 Evergreen However. stored in a retrieval system. when data is re-partitioned. All rights reserved.

2006 75 of 179 © 2006 IBM Information Integration Solutions. transmitted. Using Remove Duplicates. using the primary key column LName (ascending) and secondary key column FName (descending) would generate the resulting Data Set: Partition 0 ID 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen ID 6 4 10 LName Dodge Ford Ford Partition 2 FName John Eleanor Eleanor Address 75 Boston Boulevard 7900 Jefferson 1100 Lakeshore Partition 1 ID 5 3 9 LName Dodge Ford Ford FName Horace Edsel Edsel Address 17840 Jefferson 7900 Jefferson 1100 Lakeshore ID 1 7 LName Ford Ford Partition 3 FName Henry Henry Address 66 Edison Avenue 4901 Evergreen Note that the partition and sort keys do not always have to match. we want to select the most recent order for a given customer. For example. secondary sort keys can be used to establish order within a group for selection with the Remove Duplicates stage (which can specify First or Last duplicate to retain). sequential ordered result is needed. No part of this publication may be reproduced. .Information Integration Solutions Center of Excellence Partition 0 ID 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen ID 4 6 10 LName Ford Dodge Ford Partition 2 FName Eleanor John Eleanor Address 7900 Jefferson 75 Boston Boulevard 1100 Lakeshore Partition 1 ID 3 5 9 LName Ford Dodge Ford FName Edsel Horace Edsel Address 7900 Jefferson 17840 Jefferson 1100 Lakeshore Partition 3 ID 1 7 LName Ford Ford FName Henry Henry Address 66 Edison Avenue 4901 Evergreen Applying a parallel sort to this partitioned input Data Set. All rights reserved. transcribed. Let’s say that an input Data Set consists of order history based on CustID and Order Date. in general it is best to use a two step process: . 6. with Duplicate To Retain=First 12.4. To satisfy these requirements we could:  Partition on CustID to group related records  Sort on OrderDate in Descending order  Remove Duplicates on CustID.2 Sorting and Hashing Advanced Example provides a more detailed discussion and example of partitioning and sorting.2 Complete (Total) Sort If a single.partition and parallel Sort on key column(s) Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system.

.3. or translated into any language in any form by any means without the written permission of IBM. 6. The stand-alone sort offers more options. but as a separate stage makes job maintenance slightly more complicated. 6. links that have sort defined will have a Sort icon in addition to the partitioning icon. when specifying a keyed partitioning method.3 Link Sort and Sort Stage DataStage Enterprise Edition provides two methods for parallel sorts – the standalone sort stage (when execution mode is set to Parallel) and sort on a link (when using a keyed input partitioning method).Sorting character set: ASCII (default) or EBCDIC (strings) Parallel Framework Red Book: Data Flow Job Design July 17. the “Perform Sort” option is checked.key column usage: sorting. When specifying key column(s) for partitioning. transmitted. or both . partitioning. (Sorting on a link is not available with Auto partitioning.sort direction: Ascending or Descending . The Link sort offers fewer options. as there are fewer stages on the design canvas. but is easier to maintain in a DataStage job.Information Integration Solutions Center of Excellence - use a Sort Merge collector on these same key column(s) to generate a sequential. as shown below: Figure 35: Link Sort icon Additional properties can be specified by right-clicking on the key column as shown in the following illustration: Key column options let the developer specify: . Most often. both methods use the same internal sort package (the tsort operator). stored in a retrieval system. Within the Designer canvas. No part of this publication may be reproduced. By default. transcribed. use the Link sort unless a specific option is needed on the stand-alone Stage. the standalone Sort stage is used to specify the Sort Key mode for partial sorts. All rights reserved. ordered result set This is similar to the way parallel database engines perform their parallel sort operations.1 Link Sort Sorting on a link is specified on the Input/Partitioning stage options. In general.case sensitivity (strings) . although the Enterprise Edition engine may insert a sort if required). 2006 76 of 179 © 2006 IBM Information Integration Solutions.

For this reason. which is significantly faster than a “UNIX” sort. 2006 77 of 179 © 2006 IBM Information Integration Solutions. the Sort Key Mode is most frequently used. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys.3.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. the following properties are not available when sorting on a link: Sort Key Mode (a particularly important performance optimization) Create Cluster Key Change Column Create Key Change Column Output Statistics Sort Utility (don’t change this!) Restrict Memory Usage Figure 37: Sort Stage options Of the options only available in the standalone Sort stage. - 6. This requires some additional overhead in the sort algorithm. It is used when resorting a sub-grouping of a previously sorted input Data Set. 6. Always specify “DataStage” Sort Utility. This “subsort” uses significantly less disk space and CPU resource. instead of performing a complete Sort. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17. Specifically.5 Sub-Sorts Within the standalone Sort stage. and can often be performed in memory (depending on the size of the new subsort groups). transmitted. . It is important to note that by default the Stable sort option is disabled for sorts on a link and Enabled with the standalone Sort stage. NOTE: The Sort Utility option is an artifact of previous releases.4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. transcribed. All rights reserved. the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization. disable Stable sort unless needed.Information Integration Solutions Center of Excellence - Position of nulls in the result set (for nullable columns) Figure 36: Specifying Link Sort options 6.

DataStage Enterprise Edition inserts sort operators as necessary to ensure correct results.p1] node3[op2. as shown in this score fragment: op1[4p] {(parallel inserted tsort operator {key={value=LastName}. key={value=FirstName}}(0)) on nodes ( node1[op2. keys with “Don’t Sort (Previously Sorted)” property must be at the top of the list. If the input data does not match the key column definition for a subsort. transcribed. the key column order for these keys must match the key columns and order defined in the previously-sorted input Data Set. the job will abort. Remove Duplicates. And. No part of this publication may be reproduced.4. Sort Aggregator). All rights reserved. while new sort keys are specified with the “Sort” key mode property. 2006 78 of 179 © 2006 IBM Information Integration Solutions. . without gaps between them. Sorts are only inserted automatically when the flow developer has not explicitly defined an input sort. as shown in the following example: Figure 38: Sort Key Mode property To successfully perform a subsort.p2] node4[op2. all key columns must still be defined in the Sort stage. transmitted.6 Automatically-Inserted Sorts By default.p3] )} Typically. Parallel Framework Red Book: Data Flow Job Design July 17. The parallel job score (see 12.2 Understanding the Parallel Job Score) can be used to identify automatically-inserted sorts. or translated into any language in any form by any means without the written permission of IBM.p0] node2[op2. Re-used sort keys are specified with the “Don’t Sort (Previously Sorted)” property.Information Integration Solutions Center of Excellence To resort based on a sub-grouping. stored in a retrieval system. Enterprise Edition inserts sorts before any stage that requires matched key values or ordered groupings of (Join. 6. Merge.

memory. the following methodology should be applied when sorting in a DataStage Enterprise Edition data flow: a) b) c) d) Start with a link sort Specify only necessary key column(s) Don’t use Stable Sort unless needed Use a stand-alone Sort stage instead of a Link sort for options that not available on a Link sort:  Sort Key Mode. each sort uses 20MB of memory per partition for its memory buffer. There are two ways to prevent Enterprise Edition from inserting an un-necessary sort: a) Insert an upstream Sort stage on each link. inserted sorts can be a significant performance impact if they are not necessary. define all sort key columns with the Sort Mode key property “Don’t Sort (Previously Sorted)” b) Set the environment variable APT_SORT_INSERTION_CHECK_ONLY. Create Cluster Key Change Column. the environment variable $APT_SORT_INSERTION_CHECK_ONLY should be set to prevent Enterprise Edition from inserting un-necessary sorts before the Join stage. 6. If the sort operation can be performed in memory (as is often the case with a subsort) then no disk I/O is performed. aborting the job if data is not in the required sort order. No part of this publication may be reproduced. 2006 79 of 179 © 2006 IBM Information Integration Solutions. To perform a sort.7 Sort Methodology Using the rules and behavior outlined in the previous section. transcribed. Output Statistics  Always specify “DataStage” Sort Utility for standalone Sort stages  Use the “Sort Key Mode=Don’t Sort (Previously Sorted)” to resort a sub-grouping of a previously-sorted input Data Set e) Be aware of automatically-inserted sorts  Set $APT_SORT_INSERTION_CHECK_ONLY to verify but not establish required sort order f) Minimize the use of sorts within a job flow g) To generate a single.8 Tuning Sort Sort is a particularly expensive task within DataStage Enterprise Edition which requires CPU. sequential ordered result set use a parallel Sort and a Sort Merge collector 6. Create Key Change Column. By default.Information Integration Solutions Center of Excellence While ensuring correct results. Revisiting the partitioning examples in Section 5. This will verify sort order but not actually perform a sort. All rights reserved. This value can be changed for each standalone Sort stage using the “Restrict Memory Usage” option (the minimum is 1MB/partition).4: Partitioning Examples. transmitted. rows in the input Data Set are read into a memory buffer on each partition. stored in a retrieval system. . and disk resources. or translated into any language in any form by any means without the written permission of IBM. On a global basis. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be Parallel Framework Red Book: Data Flow Job Design July 17.

scratch disks defined in the current configuration file default disk pool . 2006 80 of 179 © 2006 IBM Information Integration Solutions. then results are temporarily spooled to disk in the following order: . Having a greater number of scratch disks for each node allows the sort to spread I/O across multiple file systems. .Information Integration Solutions Center of Excellence use to specify the size of the memory buffer. transcribed. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. If the input Data Set cannot fit into the sort memory buffer. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. stored in a retrieval system.the default directory specified by the environment variable TMPDIR . for all sort operators (link and standalone).scratch disks defined in the current configuration file (APT_CONFIG_FILE) in the “sort” named disk pool . in MB. overriding any per-sort specifications.the directory “/tmp” (on UNIX) or “C:/TMP” (on Windows) if available The file system configuration and number of scratch disks defined in parallel configuration file can greatly impact the I/O performance of a parallel sort.

transcribed. As such.Information Integration Solutions Center of Excellence 7 File Stage Usage 7. Can only be written – contents cannot be read or verified. Limitations Cannot write to a single file in parallel. but this is not recommended as it imposes risks for failure recovery. transmitted. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. Need to share data with an external Parallel SAS application. 7. Requires Parallel SAS. However. Slightly higher overhead than Data Set. Some stages (parallel Data Set) support “Append” to add new records to an existing file.2 Data Set Usage Parallel Data Sets are the persistent (on-disk) representation of the in-memory data structures of DS/EE.1 Which File Stage to Use DataStage/EE offers various stages for reading from and writing to files. Can only be read from and written to by DataStage parallel jobs or orchadmin command. All rights reserved. If data is Parallel Framework Red Book: Data Flow Job Design July 17. Cannot write in parallel. Can only be used as reference link on a Lookup stage. using the internal format of the parallel engine. Complex Flat File Data Set File Set SAS Parallel Data Set Lookup File Set No DS/EE file stage supports “update” of existing records. Need to share information with external applications. Data Sets provide maximum performance for reading and writing data from disk. No part of this publication may be reproduced.) Rare instances where lookup reference data is required by multiple jobs and is not updated frequently. Recommendations for when to use a particular stage. (Requires SAS connectivity license for DataStage. such as mainframe sources with COBOL copybook file definitions. can only be read from / written to by DS/EE or Parallel SAS. can write in parallel (generates multiple segment files). performance penalty of conversion. . as no overhead is needed to translate data to the internal DS/EE representation. Sequential File Need to read source data in complex (hierarchical) format. In general. and any limitations. 2006 81 of 179 © 2006 IBM Information Integration Solutions. Data Sets store data in partitioned form. Intermediate storage between DataStage parallel jobs. are summarized below: File Stage Recommended Usage Read and write standard files in a single format. Data Sets can only be read from and written to using a DataStage parallel job. performance penalty of format conversion. does not support hierarchical data files.

stored in a retrieval system. Readers Per Node option greater than 1 useful for SMP configurations . transmitted. the Sequential File stage can only read and write data that is in flattened (row/column) format. 2006 82 of 179 © 2006 IBM Information Integration Solutions.3 Sequential File Stages (Import and Export) The Sequential File stage can be used to read from or write to one or more flat files of the same format. the degree of parallelism of the write will correspond to the number of file names specified.3 Separating I/O from Column Import If the Sequential File input cannot be read in parallel.2 Writing to a Sequential File in Parallel It is only possible to write in parallel from a Sequential File stage when more than one output file is specified. or translated into any language in any form by any means without the written permission of IBM.1 Reading from a Sequential File in Parallel The ability to read Sequential File(s) in parallel within Enterprise Edition depends on the Read Method and the options specified: Sequential File – options to read sequentially: Read Method: Specific Files. only one file specified.3. Parallel Framework Red Book: Data Flow Job Design July 17. As shown in the following Job fragment.Information Integration Solutions Center of Excellence 7. using the format options specified in the FileSet stage. transcribed.file may be either fixed or variable-width Read Method: Specific Files. The FileSet stage will write in parallel. in parallel. All rights reserved. 7. . A better option for writing to a set of Sequential Files in parallel is to use the FileSet stage. This will create a single header file (in text format) and corresponding data files.3. performance can still be improved by separating the file I/O from the column parsing operation. define a single large string column for the non-parallel Sequential File read. input row order is not maintained across readers. No part of this publication may be reproduced. only one file specified may be a file or named pipe Read Method: File Pattern Sequential File – options to read in parallel: Read Method: Specific Files. set environment variable $APT_IMPORT_PATTERN_USES_FILESET Read Method: Specific Files. The formatting and column properties of the Column Import stage match those of the Sequential File stage.file may only be fixed-width Note that when reading in parallel. Read From Multiple Nodes option is set to Yes useful for cluster and Grid configurations . 7. In these instances.3. 7. and then pass this to a Column Import stage to parse the file in parallel. more than one file specified each file specified within a single Sequential File stage must be of the same format Read Method: File Pattern. Unlike the Complex Flat File stage.

Information Integration Solutions Center of Excellence Figure 39: Column Import example Note that this method is also useful for External Source and FTP sequential source stages. transcribed. It is also important to remember that this setting will apply to all Sequential File stages in the data flow. 7.5 Sequential File (Export) Buffering By default. or by using a File Pattern). Setting this value to a low number (such as 1) is useful for realtime applications. It is important to use ROUNDROBIN partitioning (or other partitioning appropriate to downstream components) to evenly distribute the data in the flow. instead of statically through Table Definitions.3. since the Import / Export properties used by the Sequential File and Column Import stages are not documented in the DataStage Parallel Job Developer’s Guide. but there is a small performance penalty associated with increased I/O. Note that this document is required. The format of the Schema File.4 Partitioning Sequential File Reads Care must be taken to choose the appropriate partitioning method from a Sequential File read: • Don’t read from Sequential File using SAME partitioning in the downstream stage! Unless more than one source file is specified. The environment variable $APT_EXPORT_FLUSH_COUNT allows the job developer to specify how frequently (in number of rows) that the Sequential File stage flushes its internal buffer on writes. 7.3. • 7. When a job completes successfully. stored in a retrieval system. Using the Schema File option allows the format of the source file to be specified at runtime. . 2006 83 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM.6 Parameterized Sequential File Format The Sequential File stage supports a Schema File option to specify the column definitions and file format of the source file.3. 7. the Sequential File (export operator) stage buffers its writes to optimize performance. making the entire downstream flow run sequentially (unless it is later repartitioned).7 Reading and Writing Nullable Columns When reading from or writing to Sequential Files or File Sets. as shown below: Parallel Framework Red Book: Data Flow Job Design July 17.3. No part of this publication may be reproduced. SAME will read the entire file into a single partition. including Sequential File import / export format properties is documented in the Orchestrate Record Schema manual. the buffers are always flushed to disk. transmitted. All rights reserved. each file’s data is read into a separate partition. When multiple files are read by a single Sequential File stage (using multiple files. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column.

All import and export properties are listed in the Import/Export Properties chapter of the Orchestrate Operators Reference. By default. Integer. Varchar). bounded-length Varchar columns (Varchars with the length option set). Decimal. Double-click on the column number in the grid dialog or right mouse click on the column and select edit column to set these properties.3. you must define the null field value and length in the Nullable section of the column property. transcribed. All rights reserved. When writing fixed-length files from variable-length fields (eg. if the source file has fields with values longer than the maximum Parallel Framework Red Book: Data Flow Job Design July 17. Varchar).Information Integration Solutions Center of Excellence Figure 40: Extended Column Metadata (Nullable properties) 7. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. Double-click on the column number in the grid dialog to set this column property. the field width column property must be set to match the fixed-width of the input column. Decimal.8 Reading from and Writing to Fixed-Length Files Particular attention must be taken when processing fixed-length fields using the Sequential File stage: • If the incoming columns are variable-length data types (for example. . transmitted. use the print_field import property. Integer. Double-click on the column number in the grid dialog to set this column property. No part of this publication may be reproduced. • • 7. Use caution when specifying this option as it can generate an enormous amount of detail in the job log. To display each field value.3. 2006 84 of 179 © 2006 IBM Information Integration Solutions. • If a field is nullable.9 Reading Bounded-Length VARCHAR Columns Care must be taken when reading delimited. the field width and pad string column properties must be set to match the fixed-width of the output column.

with a default of 128 (128K). 7. All rights reserved.Information Integration Solutions Center of Excellence Varchar length. setting the environment variable $APT_CONSISTENT_BUFFERIO_SIZE to a value equal to the read/write size in bytes can significantly improve performance of Sequential File operations.) When used as a target.10 Tuning Sequential File Performance On heavily-loaded file servers or some RAID/SAN array configurations. Increasing this size may improve performance. the stage allows you to write data to one or more complex flat files. transmitted. It does not write to MVS datasets. transcribed. 7. but you can configure the stage to execute sequentially if it is only reading one file with a single reader.4 Complex Flat File Stage The Complex Flat File (CFF) stage can be used to read or write one or more files in the same hierarchical format. including MVS datasets with QSAM and VSAM files. or OCCURS clauses. stored in a retrieval system. NOTE: The Complex Flat File stage cannot read from sources with OCCURS DEPENDING ON clauses. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 85 of 179 © 2006 IBM Information Integration Solutions.3. or translated into any language in any form by any means without the written permission of IBM. these extra characters will be silently truncated. Finally. When used as a source. These settings specify the size of the read (import) and write (export) buffer size in Kbytes. The environment variable $APT_IMPORT_REJECT_STRING_FIELD_OVERRUNS will direct Enterprise Edition to reject records with strings longer than their declared maximum column length. REDEFINES. (This is an error in the DataStage documentation. Complex Flat File source stages execute in parallel mode when they are used to read multiple files. A complex flat file may contain one or more GROUPs. in some disk array configurations. the stage allows you to read data from one or more complex flat files. the environment variables $APT_IMPORT_BUFFER_SIZE and $APT_EXPORT_BUFFER_SIZE can be used to improve I/O performance. .

trailing separate.y] decimal[x+y. native binary character character for filler varchar decimal decimal display_numeric display_numeric display_numeric display_numeric display_numeric display_numeric float float graphic_n. or translated into any language in any form by any means without the written permission of IBM. native binary binary. stored in a retrieval system.1 CFF Stage Data Type Mapping When you work with mainframe data using the CFF stage.y] decimal[x+y. native binary binary. native binary binary.y] sfloat dfloat ustring[n] ustring[max=n] subrec Internal Options packed packed zoned zoned.4. transmitted. leading Parallel Framework Red Book: Data Flow Job Design July 17. the data types are mapped to internal Enterprise Edition data types as follows: COBOL Type S9(1-4) COMP/COMP-5 S9(5-9) COMP/COMP-5 S9(10-18) COMP/COMP-5 9(1-4) COMP/COMP-5 9(5-9) COMP/COMP-5 9(10-18) COMP/COMP-5 X(n) X(n) X(n) 9(x)V9(y)COMP-3 S9(x)V9(y)COMP-3 9(x)V9(y) S9(x)V9(y) S9(x)V9(y) SIGN IS TRAILING S9(x)V9(y) SIGN IS LEADING S9(x)V9(y) SIGN IS TRAILING SEPARATE S9(x)V9(y) SIGN IS LEADING SEPARATE COMP-1 COMP-2 N(n) or G(n) DISPLAY-1 N(n) or G(n) DISPLAY-1 Group Description binary. All rights reserved.Information Integration Solutions Center of Excellence 7. leading separate. .y] decimal[x+y. 2006 86 of 179 © 2006 IBM Information Integration Solutions.y] or string[x+y] decimal[x+y.y] decimal[x+y.y] decimal[x+y. native binary binary. transcribed. No part of this publication may be reproduced. trailing zoned. trailing zoned. graphic_g vargraphic_g/n Size 2 bytes 4 bytes 2 bytes 2 bytes 4 bytes 8 bytes n bytes n bytes n bytes (x+y)/2+1 bytes (x+y)/2+1 bytes x+y bytes x+y bytes x+y bytes x+y bytes x+y+1 bytes x+y+1 bytes 4 bytes 8 bytes n*2 bytes n*2 bytes Internal Type int16 int32 int64 uint16 uint32 uint64 string(n) raw(n) string(max=n) decimal[x+y. native binary binary.y] or string[x+y] decimal[x+y.

entries are placed in the Director job log.1. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17. For this reason. it is important to minimize the number of transformers. For example. This makes it easy to identify reject conditions (by row counts). it should be converted from its out-of-band (internal) null representation to an in-band (specific value) null representation using stage variables or the Modify stage. by definition. Because the DS/EE Transformer runs in parallel. 2006 87 of 179 © 2006 IBM Information Integration Solutions. and to use other stages (such as Copy) when derivations are not needed. Always include reject links in a parallel Transformer. To create a Transformer reject link in Designer. 8. See Section 3. if you intend to use a nullable column within a Transformer derivation or output link constraint. All rights reserved. No part of this publication may be reproduced. the Transformer will reject the row instead of sending it to the output link(s). the Transformer will allow this row to be sent to the output link without being rejected. When rows are rejected by a Transformer. stored in a retrieval system.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints.1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component.1. transmitted.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer. right-click on an output link and choose “Convert to Reject: Figure 41: Transformer Reject link The parallel Transformer rejects NULL derivation results (including output link constraints) because the rules for arithmetic and string handling of NULL values are.col) Then “” Else link.col Note that if an incoming column is only used in an output column mapping. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link.8.4: Parallel Transformer stages for guidelines on Transformer stage usage. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. 8. transcribed. For this reason. . undefined.Information Integration Solutions Center of Excellence 8 Transformation Languages 8. Even if the target column in an output derivation allows nullable results. or translated into any language in any form by any means without the written permission of IBM.

the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. The environment variable APT_DECIMAL_INTERM_ROUND_MODE can be used to change the rounding behavior using one of the following keywords: ceil Rounds towards positive infinity. not the target. Since the Transformer will abort the entire job flow immediately. Examples: 1. transcribed.6 -> -1 floor Rounds towards negative infinity. Thus. -1. No part of this publication may be reproduced. Create a new output link that will handle rows that match the abort rule. -1. 8.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition.4 -> -1. stored in a retrieval system. For example. When generating a sequence of numbers in parallel.1. or translated into any language in any form by any means without the written permission of IBM.4 -> 2.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment. the incoming column must be type VarChar before it is evaluated in the Transformer. internal decimal results are rounded to the nearest applicable value. 1.4 -> 1. 1). or committed to database tables. TrimLeadingTrailing(string) works only if string is a VarChar field. Within the link constraints dialog box. these internal decimal variables will have a precision and scale of 38 and 10. For example.5-> 2. and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example. or performing parallel derivations.4 -> -2 round_inf Rounds or truncates towards nearest representable value. Examples: 1. Therefore. By default. Examples: 1.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage. the PadString function uses the length of the source type. All rights reserved. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7.4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule. there are times when internal decimal variables need to be generated in order to perform the evaluation. up to a maximum precision of 255 and scale of 125. If more precision is required. apply the abort rule to this output link.1. 8. . -1. it is important to make sure the type conversion is done before a row reaches the Transformer.1.3.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17. 8. -1.5: Sequential File (Export) Buffering).6 ->1. 2006 88 of 179 © 2006 IBM Information Integration Solutions. By default. the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range. it is possible that valid rows will not have been flushed from Sequential File (export) buffers. transmitted. breaking ties by rounding positive values toward positive infinity and negative values toward negative infinity.

the stage variable definition would be: Parallel Framework Red Book: Data Flow Job Design July 17. unless the derivation is empty For each output link: Evaluate the link constraint. By doing this. All rights reserved. 8.5..1. From this sequence..3] is evaluated for each column that uses it.5. . For example. suppose multiple columns in output links want to use the same substring of an input column. the output links are also evaluated in the order in which they are displayed. 2006 89 of 179 © 2006 IBM Information Integration Solutions. it can be seen that there are certain constructs that would be inefficient to include in output column derivations.56 -> 1. round or truncate to the scale size. then the following test may appear in a number of output columns derivations: IF (DSLINK1.56 ->-1. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example. The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value. This can be made more efficient by moving the substring calculation into a stage variable. or translated into any language in any form by any means without the written permission of IBM. the evaluation of the substring of DSLINK1.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations. transmitted. the substring is evaluated just once for every input row. In this case. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. No part of this publication may be reproduced. transcribed.3] = “001”) THEN . Examples: 1. as they would be evaluated once for every output column that uses them. it is useful to understand what items get evaluated and when. In this case. Similarly.col[1. if true Evaluate each column derivation value Write the output record Else skip the link Next output link Next input row The stage variables and the columns within a link are evaluated in the order in which they are displayed in the Transformer editor. stored in a retrieval system.col[1. -1.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign.

. Any expression that previously used this function would be changed to use the stage variable instead. In this case. Therefore. this example could be improved further by also moving the string comparison into the stage variable. the variable would have its initial value set to: Str(“ “. it is not reevaluated for each input row. . transcribed. its value for the whole Transformer processing is unchanged from the initial value. The solution here is to move the function evaluation into the initial value of a stage variable. stored in a retrieval system. a column definition may include a function call that returns a constant value. In this case. All rights reserved.20) This returns a string of 20 spaces. In fact. before any input rows are processed.Information Integration Solutions Center of Excellence DSLINK1. such as: Str(“ “. the function would still be evaluated once for every input row.20) You would then leave the derivation of the stage variable on the main Transformer page empty.col[1. or translated into any language in any form by any means without the written permission of IBM. the function would be evaluated every time the column derivation is evaluated.. A stage variable can be assigned an initial value from the Stage Properties dialog/Variables tab in the Transformer stage editor.3] = “001” THEN 1 ELSE 0 and each column derivation would start with: IF (StageVar1) THEN This reduces both the number of substring functions evaluated and string comparisons made in the Transformer. but in this case.col1[1. It would be more efficient to calculate the constant value just once for the whole Transformer. another example would be part of an expression such as: Parallel Framework Red Book: Data Flow Job Design July 17. 2006 90 of 179 © 2006 IBM Information Integration Solutions. The initial value of the stage variable is evaluated just once. transmitted. because the derivation expression of the stage variable is empty. In addition to a function value returning a constant value. This can be achieved using stage variables. Then.3] and each column derivation would start with: IF (StageVar1 = “001” THEN . This function could be moved into a stage variable derivation. • Where an expression includes calculated constant values For example. No part of this publication may be reproduced. The stage variable would be: IF (DSLink1.

then. where that conversion would have been required. As noted in the previous section. stored in a retrieval system. The Modify stage uses the syntax of the underlying modify operator. The solution in this case is just to change the constant from a string to an integer: DSLink1.col1+"1" In this case. since it uses low-level functionality that is part of every DataStage Enterprise Edition component. an integer stage variable. the data type of the stage variable should be set correctly for that context. documented in the Parallel Job Developers Guide as well as the Orchestrate Operators Reference. the "1" is a string constant. null conversion. if an input column is used in more than one expression.col1 were a string field.col1. transmitted. For example. a conversion would be required every time the expression is evaluated. • Where an expression requiring a type conversion is used as a constant. Since the subpart of the expression is actually constant. dropping and renaming columns. you would create. then it would be more efficient to use a stage variable to perform the conversion once. needless conversions are required wherever that variable is used. In this case. using the initial value setting to perform the concatenation just once.2 Modify Stage The Modify stage is the most efficient “stage” available. All rights reserved. No part of this publication may be reproduced. where it requires the same type conversion in each expression. and then use the stage variable in place of DSLink1. again. in order to be able to add it to DSLink1. if DSLINK1.col1+1 In this example.Information Integration Solutions Center of Excellence "abc" : "def" As with the function-call example. or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. Otherwise. 8. It should be noted that when using stage variables to evaluate parts of expressions. for example. . The standalone Modify stage can be used for non-default type conversions (nearly all date and time conversions are non-default). the Output Mapping properties for any parallel stage will generate an underlying modify for default data type conversions.col1. this concatenation is evaluated every time the column derivation is evaluated. 2006 91 of 179 © 2006 IBM Information Integration Solutions. it must be converted from a string to an integer each time the expression is evaluated. an expression may include something like this: DSLink1. However. this constant part of the expression could again be moved into a stage variable. and so.col1. this would be fine. transcribed. and string trim. or it is used in multiple places. If this just appeared once in one output column expression. specify its derivation to be DSLINK1.

use it if you are also converting types. For a numeric field value can be a numeric value. and timestamp fields. NOTE: The DataStage Parallel Job Developers Guide gives incorrect syntax for converting an out-of-band null to an in-band null (value) representation. 8.Information Integration Solutions Center of Excellence 8.dataType is its optional data type.2. The following example removes all leading ASCII NULL characters from the beginning of name and places the remaining characters in an output variable-length string with the same name: Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. transmitted. value can be a string. the syntax is: destField[:dataType] = make_null(sourceField. the syntax is: destField[:dataType] = handle_null (sourceField. direction defaults to end. with the following syntax: stringField=string_trim[character. To convert from an out-of-band null to an in-band null (value) representation within Modify. All rights reserved. 2006 92 of 179 © 2006 IBM Information Integration Solutions.sourceField is the source field’s name.dataType is its optional data type. By default. To convert from an in-band null to an out-of-band null. Justify has no affect when the target string has variable length. string. this is NULL. The destField is converted from an Orchestrate out-of-band null to a value of the field’s data type. .2 Modify and String Trim The function string_trim has been added to Modify. .value) where: . The value of the direction and justify arguments can be either begin or end. The character argument is the character to remove. .1 Modify and Null Handling The Modify stage can be used to convert an out-of-band null value to an in-band null representation and vice-versa. date. use it if you are also converting types.destField is the destination field’s name.value is the value you wish to represent a null in the output. or translated into any language in any form by any means without the written permission of IBM. justify] (string) You can use this function to remove the characters used to pad variable-length strings when they are converted to fixed-length strings of greater length. No part of this publication may be reproduced.sourceField is the source field’s name . . stored in a retrieval system.value is the value of the source field when it is null.destField is the destination field’s name.2. time.value) where: . . these characters are retained when the fixed-length string is then converted back to a variable-length string. . for decimal. direction. and justify defaults to begin. By default.

Information Integration Solutions Center of Excellence name:string = string_trim[NULL. transcribed. or translated into any language in any form by any means without the written permission of IBM. transmitted. end. and left-justifies the resulting hue fixed-length string: hue:string[10] = string_trim[‘Z’. . stored in a retrieval system. All rights reserved. begin](color) Parallel Framework Red Book: Data Flow Job Design July 17. 2006 93 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. begin](name) The following example removes all trailing Z characters from color.

when a match does not occur. all rows on an outer link (for example. Join vs. to each of the inner links and test that column for the constant after you have performed the join. For this reason. the default value for a Varchar is an empty string (“”). If un-matched rows must be captured or logged.1 Lookup vs. 9. or translated into any language in any form by any means without the written permission of IBM. the Join stage inserts values into the unmatched non-key column(s) using the following rules: a) If the non-key column is defined as nullable (on the Join input links) then Enterprise Edition will insert NULL values in the unmatched columns b) If the non-key column is defined as not-nullable. with a constant value. the default value for an Integer is zero. In an OUTER join scenario. it is best to use a Column Generator to add an ‘indicator’ column. and to retest if the relative volumes of data change dramatically. Oracle Enterprise. In most cases. (see Section 10.Information Integration Solutions Center of Excellence 9 Combining Data 9. For example. If the Data Sets are larger than available memory resources.7: Database Sparse Lookup vs. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. No part of this publication may be reproduced.1. stored in a retrieval system. transmitted. During an Outer Join. the JOIN or MERGE stage should be used. This isolates your match/no-match logic from any changes in the metadata. 2006 94 of 179 © 2006 IBM Information Integration Solutions. Sparse Lookups may also be appropriate for exception-based processing when the number of exceptions is a small fraction of the main input data. This is also handy with Lookups that have multiple reference links. Limit the use of database Sparse Lookups (available in the DB2 Enterprise. All rights reserved. Each lookup reference requires a contiguous block of shared memory. and the default value for a Char is a string of padchar characters equal to the length of the Char column. Join). It is best to test both the Sparse and Normal to see which actually performs best. Right Outer. care must be taken to change the column properties to allow NULL values before the Join. Left Outer. and ODBC Enterprise stages) to scenarios where the number of input rows is significantly smaller (for example. This is most easily done by inserting a Copy stage and mapping a column from NON-NULLABLE to NULLABLE. A Transformer stage can be used to test for NULL values in unmatched columns. or both links in the case of Full Outer) are output regardless of match on key values. Merge The Lookup stage is most appropriate when the reference data for all lookup stages in a job is small enough to fit into available physical memory. 1:100 or more) than the number of reference rows. then Enterprise Edition inserts “default” values based on the data type. an OUTER join operation must be performed. .2 Capturing Unmatched Records from a Join The Join stage does not provide reject handling for unmatched records (such as in an InnerJoin scenario).

To perform a total aggregation. transmitted. which maintains the results of each keycolumn value/aggregation pair in memory. the Hash Aggregator should only be used when the number of distinct key values is small and finite. stored in a retrieval system. 9. Because each key value/aggregation requires approximately 2K of memory. The Sort Aggregation Method should be used when the number of key values is unknown or very large. All rights reserved. .3. the default Aggregation Method is set to Hash. or translated into any language in any form by any means without the written permission of IBM. aggregate on generated key column) there is no need to sort or hash-partition the input data with only one key column value aggregate Sequentially on the generated column - - Note that in this example use two Aggregators are used to prevent the sequential aggregation from disrupting upstream processing. the Sort Aggregator requires presorted data.Information Integration Solutions Center of Excellence 9. set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. Parallel Framework Red Book: Data Flow Job Design July 17. To aggregate in decimal precision. An exception to this is financial calculations which should be done in decimal to preserve appropriate precision. 9. No part of this publication may be reproduced. transcribed. 2006 95 of 179 © 2006 IBM Information Integration Solutions. Unlike the Hash Aggregator. Note that performance is typically better if you let calculations occur in floating point (Double) data type and convert the results to decimal downstream in the flow.3.3 The Aggregator Stage 9.3. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property. but only maintains the calculations for the current group in memory. the output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double).1 Aggregation Method By default.2 Aggregation Data Type By default.3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings. use the stages shown on the right to: generate a single constant-value key column using the Column Generator or an upstream Transformer aggregate in parallel on the generated column (partition Round Robin.

1. You may need to customize the palette to add hidden stages. transmitted. Plug-In Database Stages Dynamic RDBMS DB2/UDB API DB2/UDB Load Informix CLI Informix Load Informix XPS Load Oracle OCI Load RedBrick Load Sybase IQ12 Load Sybase OC Teradata API Teradata MultiLoad (MultiLoad) Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. Oracle. No part of this publication may be reproduced.1 Database stage types DataStage Enterprise Edition offers database connectivity through native parallel and plug-in stage types. and Teradata). For some databases (DB2. Teradata API) are visible in the default DataStage Designer palette. multiple stage types are available: Teradata MultiLoad (TPump) Native Parallel Database Stages DB2/UDB Enterprise Informix Enterprise ODBC Enterprise Oracle Enterprise SQL Server Enterprise Teradata Enterprise NOTE: Not all database stages (for example. 2006 96 of 179 © 2006 IBM Information Integration Solutions. Informix.Information Integration Solutions Center of Excellence 10 Database Stage Guidelines 10. or translated into any language in any form by any means without the written permission of IBM.1 Database development overview This section is intended to provide guidelines appropriate to accessing any database within DataStage Enterprise Edition. . Subsequent sections provide database-specific tips and guidelines. transcribed. 10. stored in a retrieval system.

From a design perspective. care must be taken to assign the correct data types in the job design.1.1. the DataDirect ODBC Driver Manager is also included. native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform.1.1.2 ODBC Enterprise stage In general. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect. However. Native parallel stages always pre-query the database for actual runtime metadata (column names. they should only be used when it is not possible to use a native parallel stage. the native parallel stages often have more stringent connectivity requirements than plug-in stages. Because of their tight integration with database technologies. or translated into any language in any form by any means without the written permission of IBM. for maximum parallel performance. On UNIX. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform. Because there are exceptions to this rule (especially with Teradata). so Table Definitions must match the order of columns in a query. The benefit of ODBC Enterprise stage comes from the large number of included and third party ODBC drivers to enable connectivity to all major database platforms. plug-in database stages match columns by order.1. . Because plug-in stage types cannot read in parallel. 10. not position in the stage Table Definitions. operating system.Information Integration Solutions Center of Excellence 10. and version. Unlike the database-specific parallel stages. No part of this publication may be reproduced. scalability. specific guidelines of when to use various stage types are provided in the database-specific topics in this section.1. transcribed. not name. the ODBC Enterprise stage cannot interface with database-specific parallel load technologies. and cannot span multiple servers in a clustered or Grid configuration. the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. transmitted. types. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. This allows Enterprise Edition to match return columns by name.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. stored in a retrieval system. attributes). 10.1 Native Parallel database stages In general. 2006 97 of 179 © 2006 IBM Information Integration Solutions. For example. the ODBC Enterprise stage cannot read in parallel (although a patch to allow parallel read may be available on some platforms through IBM IIS Support). Furthermore. and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied.

However. 2006 98 of 179 © 2006 IBM Information Integration Solutions.2. use orchdbutil to import metadata to avoid type conversion issues.1.both DS/EE and relational databases support null values. For each native parallel database stage: . 10.2 Database Metadata 10.1. and selecting “Import from Database Table” option in the wizard as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. Database-specific data type mapping tables are included in the following sections. data types. All rights reserved.1 Runtime metadata At runtime.columns of the database row correspond to columns of a DS/EE record .2 Metadata Import When using the native parallel DB2 Enterprise. nullability) and partitioning scheme (in some cases) of the source or target table. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column The actual metadata used by a DS/EE native parallel database stage is always determined at runtime. care must be taken that the column data types defined by the DataStage developer match the data types generated by the database stage at runtime. This allows the database stages to match return values by column name instead of position. the DS/EE native parallel database stages always “pre-query” the database source or target to determine the actual metadata (column names.Information Integration Solutions Center of Excellence 10. or translated into any language in any form by any means without the written permission of IBM.2. This utility is available as a server command line utility and within Designer and Manager using “Import Orchestrate Schema Definitions”. regardless of the table definitions assigned by the DataStage developer. stored in a retrieval system.1. . transcribed.rows of the database result set correspond to records of a DS/EE Data Set . No part of this publication may be reproduced. Informix Enterprise or Oracle Enterprise stages.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types .

there may be cases where user-defined functions or logic need to be executed on the database server. or translated into any language in any form by any means without the written permission of IBM. similar to the illustration on the right: Parallel Framework Red Book: Data Flow Job Design July 17. it is best to explicitly specify column names on all source database stages. the DataStage Designer will automatically populate the select list based on the stage’s output column definition.2. . The only exception to this rule is when building dynamic database jobs that use runtime column propagation to process all columns in a source table. For example. transcribed. For “Auto-Generated” SQL.1. SUM(sales) Total FROM store_info GROUP BY store_name Note that in many cases it may be more appropriate to aggregate using the Enterprise Edition Aggregator stage. Test the connection using View Data button. No part of this publication may be reproduced. orchdbutil can be scripted to automate the process of importing a large number of tables. always specify the “Select List” subproperty. All rights reserved. 10. 2006 99 of 179 © 2006 IBM Information Integration Solutions. you will see a window with the result columns and data. stored in a retrieval system.1. it is important to use SQL aliases to explicitly name the calculated columns so that they can be referenced within the DataStage job. As a command. 10. it will be easier to use the corresponding orchdbutil command-line utility from the DataStage server machine. the following SQL assigns the alias Total to the calculated column: SELECT store_name. transmitted.3 Optimizing Select Lists For best performance and optimal memory usage. The alias name(s) should then be added to the Table Definition within DataStage. 10. When importing a large number of tables.4 Testing Database Connectivity The “View Data” button on the Output / Properties tab of source database stages lets you verify database connectivity and settings without having to create and run a job. instead of using an unqualified “Table” or SQL “SELECT *” read. If the connection is successful.Information Integration Solutions Center of Excellence Figure 42: orchdbutil metadata import One disadvantage to the graphical orchdbutil metadata import is that the user interface requires each table to be imported individually.1. However. For “Table” read method.3 Defining Metadata for Database Functions When using database functions within a SQL SELECT list in a Read or Lookup.

it is not a good idea to let DataStage generate target tables unless they are used for temporary storage. etc) not possible with the “Create” option. transcribed.5 Designing for Restart To enable restart of high-volume jobs. or translated into any language in any form by any means without the written permission of IBM. and doing so may violate data-management (DBA) policies. These options allow commands (including SQL) to be sent to the database before (OPEN) or after (CLOSE) all rows are read/written/loaded to the database. and the CLOSE command could be used to select all rows from the temporary table and insert into a final target table. For example. logging. the OPEN command could be used to create a temporary table. Parallel Framework Red Book: Data Flow Job Design July 17. .1. a Lookup stage (or Join stage. constraints. 2006 100 of 179 © 2006 IBM Information Integration Solutions. Subsequent job(s) should read this Data Set and populate the target table using the appropriate database stage and write method. There are limited capabilities to specify Create table options in the stage. transmitted. Figure 44: View Additional Error Detail 10. it is important to separate the transformation process from the database write (Load or Upsert) operation. including databasespecific options (tablespace. As another example. the OPEN command can be used to create a target table. and you will be prompted to view additional detail. All rights reserved. OPEN and CLOSE are not offered by plugin database stages.1. In general. No part of this publication may be reproduced. After transformation. As a further optimization. the results should be landed to a parallel Data Set. Clicking YES will display a detailed dialog box with the specific error messages generated by the database stage that can be very useful in debugging a database connection failure.6 Database OPEN and CLOSE Commands The native parallel database stages provide options for specifying OPEN and CLOSE commands. an error message may appear. depending on data volume) can be used to identify existing rows before they are inserted into the target table. stored in a retrieval system.Information Integration Solutions Center of Excellence Figure 43: Sample View Data Output If the connection fails. 10.

All rights reserved. transcribed. Join Data read by any database stage can serve as the reference input to a Lookup operation. and leverages the database capabilities. when reading from DB2.1. 10. metadata capture and re-use. a Sparse Lookup may be appropriate. IMPORTANT: The individual SQL statements required by a “Sparse” Lookup are an expensive operation from a performance perspective. this lock is not sent – and should be specified explicitly if appropriate. Parallel Framework Red Book: Data Flow Job Design July 17. For scenarios where the number of input rows is significantly smaller (for example. 2006 101 of 179 © 2006 IBM Information Integration Solutions. and Oracle Enterprise stages allow the lookup type to be changed to “Sparse”.Information Integration Solutions Center of Excellence It is important to understand the implications of specifying a user-defined OPEN and CLOSE command. 1:100 or more) than the number of reference rows in a DB2 or Oracle table. For example. 10. the DB2/UDB Enterprise. While there are extreme scenarios when the appropriate technology choice is clearly understood.8 Appropriate Use of SQL and DataStage When using relational database sources. the optimal implementation involves leveraging the strengths of each technology to provide maximum throughput and developer productivity. sending individual SQL statements to the reference database for each incoming Lookup row. a default OPEN statement places a shared lock on the source. When directly connected as the reference link to a Lookup stage. In most cases. transmitted. . there is often a functional overlap between SQL and DataStage functionality. When specifying a user-defined OPEN command. or translated into any language in any form by any means without the written permission of IBM. This minimizes impact on network and memory resources. this reference data is loaded into memory like any other reference link (“Normal” Lookup). ODBC Enterprise. with no intermediate stages. By default. stored in a retrieval system. The following guidelines can assist with the appropriate use of SQL and DataStage technologies in a given job flow: • When possible. use a SQL filter (WHERE clause) to limit the number of rows sent to the DataStage job. and ongoing application maintenance costs. Further details are outlined in the respective database sections of the Orchestrate Operators Reference which is part of the Orchestrate OEM documentation. No part of this publication may be reproduced. there may be “gray areas” where the decision should be made based on factors such as developer productivity. Sparse Lookup is only available when the database stage is directly connected to the reference link.7 Database Sparse Lookup vs.1. Although it is possible to use either SQL or DataStage to solve a given business problem. it is faster to use a DataStage JOIN stage between the input and DB2 reference data than it is to perform a “Sparse” Lookup.

• • Parallel Framework Red Book: Data Flow Job Design July 17. A join that reduces the result set significantly is also often appropriate to do in the database. the efficiency of the Enterprise Edition Sort and Join stages can be significantly faster than an equivalent SQL query. or translated into any language in any form by any means without the written permission of IBM. or when the source includes a large number of database tables. especially when the join columns are indexed. Avoid the use of database stored procedures (for example. transcribed. 2006 102 of 179 © 2006 IBM Information Integration Solutions. it can still be beneficial to use database filters (WHERE clause) if appropriate. stored in a retrieval system. it is best to implement business rules using native parallel DataStage components. Oracle PL/SQL) on a per-row basis within a high-volume data flow. When combining data from very large tables. All rights reserved.Information Integration Solutions Center of Excellence • Use a SQL Join to combine data from tables with a small number of rows in the same database instance. In this scenario. . For maximum scalability and parallel performance. transmitted. No part of this publication may be reproduced.

2006 103 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence 10. Parallel Framework Red Book: Data Flow Job Design July 17. Check with IBM IIS Support for availability. No part of this publication may be reproduced. (Before DB2 v8. This will only work when the authentication mode of the database on the remote instance is set to “client authentication”. If you use the stage in this way. For this reason.2 DB2 Guidelines 10.1 DB2 Stage Types DataStage Enterprise Edition provides access to DB2 databases using one of 5 stages. and DataStage Plug-In guides) 10. lookup. the “client authentication” configuration of a remote instance is not recommended. the DB2 hardware/UNIX/software platform must match the hardware/software platform of the DataStage ETL server. this was also called “DB2 EEE”. you may experience data duplication when working in partitioned instances since the node configuration of the local instance may not be the same as the remote instance. . upsert. All rights reserved. The DB2/UDB Enterprise stage requires DB2 Enterprise Server Edition on UNIX with Data Partitioning Facility (DPF) option. or translated into any language in any form by any means without the written permission of IBM.) Furthermore. summarized in the following table: DataStage Stage Name DB2/UDB Enterprise DB2/UDB API DB2/UDB Load ODBC Enterprise Dynamic RDBMS Stage Type Native Parallel Plug-In Plug-In Native Plug-In DB2 Requirement DPF. transcribed.1. 3 A patched version of the ODBC Enterprise stage allowing parallel read is available from IBM IIS Support for some platforms.2. consult the DataStage documentation (DataStage Parallel Job Developers Guide. same platform as ETL server 2 Any DB2 via DB2 Client or DB2-Connect Subject to DB2 Loader Limitations Any DB2 via DB2 Client or DBE-Connect Any DB2 via DB2 Client or DB2-Connect Supports Partitioned DB2? Yes / directly to each DB2 node Yes / through DB2 node 0 No Yes / through DB2 node 0 Yes / through DB2 node 0 Parallel Read? Yes Parallel Write? Yes Parallel Sparse Lookup Yes SQL Open / Close Yes No No No3 No Possible Limitations No No Possible Limitations No No No No No No No No For specific details on the stage capabilities. 2 It is possible to connect the DB2 UDB stage to a remote database by simply cataloging the remote database in the local instance and then using it as if it were a local database. and load capabilities to parallel DB2 databases on UNIX using the native parallel DB2/UDB Enterprise stage.2.1 DB2/UDB Enterprise stage Enterprise Edition provides native parallel read. stored in a retrieval system. transmitted.

parallel component the DB2/UDB Enterprise stage is designed for maximum performance and scalability. Add the following properties: Parallel Framework Red Book: Data Flow Job Design July 17.1. using the same data partitioning as the referenced DB2 tables. Sparse Lookup is not supported through the DB2/API stage. the ability to write in parallel may be limited by the table and index configuration set by the D2 database administrator.2. The DB2/API (plug-in) stage should only be used to read from and write to DB2 databases on nonUNIX platforms (such as mainframe editions through DB2-Connect). stored in a retrieval system. 10. By facilitating flexible connectivity to multiple types of remote DB2 database servers. and reading from . databases on UNIX platforms that differ from the platform of the DataStage ETL server. Furthermore. These stages also provide connectivity to non-UNIX DB2 databases. since the DS/EE framework will instantiate multiple copies of these stages to handle the data that has already been partitioned in the parallel framework. when used as data sources.2. Because each plug-in invocation will open a separate connection to the same target DB2 database table. 2006 104 of 179 © 2006 IBM Information Integration Solutions.writing to DB2 in parallel (where appropriate). transcribed. 10. . No part of this publication may be reproduced. including direct communication with each DB2 database node.2 Connecting to DB2 with the DB2/UDB Enterprise Stage Create a Parallel job and add a DB2/UDB Enterprise stage. or DB2 databases on Windows or Mainframe platforms (except for the “Load” stage against a mainframe DB2 instance which is not supported). These goals are achieved through tight integration with the DB2 RDBMS. “DB2 EE”). All rights reserved. it may be possible to write to a DB2 target in parallel.Information Integration Solutions Center of Excellence As a native. the use of DataStage plug-in stages will limit overall performance and scalability.2 ODBC and DB2 Plug-In Stages The ODBC Enterprise and plug-in stages are designed for lower-volume access to DB2 databases without the DPF option installed (prior to v8. or translated into any language in any form by any means without the written permission of IBM. Using the DB2/UDB API stage or the Dynamic RDBMS stage. plug-in stages cannot read from DB2 in parallel. transmitted.

stored in a retrieval system. Single stream . Set this to the DB2 client instance name. Enter the user name for connecting to DB2. Optionally set this to the instance name of the DB2 server. • Client Alias DB Name. transcribed. No part of this publication may be reproduced. this is required for a remote connection in order to retrieve the catalog information from the local instance of DB2 and thus must have privileges for that local instance. • Server. DB2INSTANCE.Two Instances Only reading from one instance and writing to another instance with no other DB2 instances (not sure how many stages of these 2 instances can be added to the canvas for this configuration for lookups) 2. Since a db2nodes. Otherwise use the environment variables $APT_DBNAME or $APT_DB2DBDFT to identify the database. it is possible to connect to more than one DB2 instance within a single job. Enter the password for connecting to DB2. you must adhere to all of the directions specified for connecting to a remote instance AND the following: • You must not set the APT_DB2INSTANCE_HOME environment variable. Optionally set this to the remote server database name. All rights reserved. Once this variable is set. This is required only if the client’s alias is different from the actual name of the remote server database. 10. you will need to set the following properties on the DB2/UDB Enterprise stage in your parallel job: • Client Instance Name. • Database. this is required for a remote connection in order to retrieve the catalog information from the local instance of DB2 and thus must have privileges for that local instance.2. Set this to the DB2 client’s alias database name for the remote DB2 server database. to identify the instance name of the DB2 server. • Password. Otherwise use the DB2 environment variable. Two Stream – One Instance per Steam reading from instance A and writing to instance A and reading from instance B and writing to instance B (not sure how many stages of these 2 instances can be added to the canvas for this configuration for lookups) 3.3 Configuring DB2 Multiple Instances in One DataStage Job Although it is not officially supported. or translated into any language in any form by any means without the written permission of IBM.cfg file can July 17. transmitted. it will try to use it for each of the connections in the job. 2006 105 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Multiple Stream with N DB2 sources with no DB2 targets reading from 1 to n DB2 instances in separate source stages with no downstream other DB2 stages In order to get this configuration to work correctly. DataStage assumes you require remote connection.Information Integration Solutions Center of Excellence Figure 45: DB2/UDB Enterprise stage properties For connection to a remote DB2/UDB instance. • User. If you set this property. Your job must meet one of the following configurations (note: the use of the word “stream” refers to a contiguous flow of one stage to another within a single job): 1. .

or translated into any language in any form by any means without the written permission of IBM. 10. All rights reserved. the Enterprise Edition column will be named “APT_37dig” .2. . if the third DB2 column is named 7dig. For example. although this incurs the overhead of Sequential File stage (corresponding export/import operators) which does not run in parallel.cfg. separate jobs can communicate through named pipes. To connect to multiple DB2 instances. if the data volumes are sufficiently small. No part of this publication may be reproduced. where column# is the number of the column.the name can contain only alphanumeric and underscore characters . but have the following restrictions: .the name must start with a letter or underscore character . and Dynamic RDBMS plug-in stages. DB2 Load.Information Integration Solutions Center of Excellence only contain information for one instance. create a sqllib subdirectory and place the remote instance’s db2nodes. Dynamic RDBMS) may be used to access data in other instances. Or.5 DB2/API stage Column Names When using the DB2/API. you must build a user on the DataStage server with the same name as the instance you are trying to connect to (the default logic for the DB2/UDB Enterprise stage is to use the instance’s home directory as defined for the UNIX user with the same name as the DB2 instance). 2006 106 of 179 © 2006 IBM Information Integration Solutions.the name is case insensitive When there is an incompatibility. Enterprise Edition converts the DB2 column name as follows: . DS will default to this directory to find the configuration file for the remote instance. the native parallel DB2/UDB Enterprise stage translates column names exactly except when a component of a DB2 column name is not compatible with Enterprise Edition column naming conventions.4 DB2/UDB Enterprise stage Column Names At runtime. DB2 plug-in stages (DB2 API. Observe the following guidelines when $DS_ENABLE_RESERVED_CHAR_CONVERT is set: Parallel Framework Red Book: Data Flow Job Design July 17. In the users UNIX home directory. Depending on platform configuration and I/O subsystem performance. set the environment variable $DS_ENABLE_RESERVED_CHAR_CONVERT if your DB2 database uses the reserved characters # or $ in column names. This converts these special characters into an internal representation that DataStage can understand. the character is replaced by two underscore characters 10. this will create problems.if the DB2 column name does not begin with a letter or underscore. stored in a retrieval system. DB2 Load. Since the APT_DB2INSTANCE_HOME is not set. landing intermediate results to a parallel Data Set.2. transcribed. which place no limit on the length of a column name.if the DB2 column name contains a character that is not alphanumeric or an underscore. we recommend using separate jobs with their respective DB2 environment variable settings. • In order for DataStage to locate the db2nodes. the string “APT__column#” (two underscores) is added to beginning of the column name.cfg there. transmitted.

. Use the “db2 describe table [table-name]” command on the DB2 client for each table to be accessed to determine if UDTs are in use. decimal[p. as shown in the following table. Alternatively. No part of this publication may be reproduced. The original names are used in generated SQL statements. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. All rights reserved. the internal column names are displayed rather than the original DB2 names both in table definitions and in the Data Browser. 10. DECIMAL[p. Once the table definition is loaded.Information Integration Solutions Center of Excellence - Avoid using the strings __035__ and __036__ in your DB2 column names (these are used as the internal representations of # and $ respectively) Import meta data using the Plug-in Meta Data Import tool. stored in a retrieval system.r) NVARCHAR(n.6 DB2/UDB Enterprise stage Data Type Mapping The DB2 database schema to be accessed must NOT have any columns with User Defined Types (UDTs).s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n. The DB2/UDB Enterprise stage converts DB2 data types to Enterprise Edition data types. transmitted.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. the result is a timestamp field.s] where p is the precision and s is the scale dfloat dfloat int32 decimal string[n] or ustring[n] string[max=n] or ustring[max=n] sfloat int32 sfloat int16 string[max=n] or ustring[max=n] IMPORTANT: DB2 data types that are not listed in the above table cannot be used in the DB2/UDB Enterprise stage. transcribed. avoid hand editing (this minimizes the risk of mistakes or confusion). or translated into any language in any form by any means without the written permission of IBM. DB2 Data Type CHAR(n) CHARACTER VARYING(n. the result is a time field. They are also used in derivations and expressions. examine the DDL for each schema to be accessed.2. 2006 107 of 179 © 2006 IBM Information Integration Solutions. however. and you should use them if entering SQL in the job yourself.r) DATE DATETIME Enterprise Edition Data Type string[n] or ustring[n] string[max=n] or ustring[max=n] date Time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component.

During the load operation. transcribed. 10. when using user-defined SQL without partitioning against large volumes of DB2 data. the DB2 Load DataStage job must be re. the DB2/UDB Enterprise stage offers the choice of SQL (insert / update / upsert / delete) or fast DB2 loader methods.run in Truncate mode to clear the load pending state. the overhead of routing information through a remote DB2 coordinator may be significant. The choice between these methods depends on required performance. forcing the DB2 Enterprise stages on each ETL server to communicate directly with their local DB2 coordinator. stored in a retrieval system. In these instances. In this configuration. That is. and the availability of new rows to other applications. transmitted.2. database log usage. the DB2 Load method places an exclusive lock on the entire DB2 tablespace into which it loads the data and no other tables in that tablespace can be accessed by other applications until the load completes. All rights reserved. DB2 Enterprise stage should not include the Client Instance Name property. Parallel Framework Red Book: Data Flow Job Design July 17. communicating directly with each database node. the contents of the table are unusable and the tablespace is left in a load pending state. Time and row-based commit intervals determine the transaction size. and recoverability. and the target table(s) may be accessed by other users. the DB2/UDB Enterprise stage is designed for maximum performance and scalability against very large partitioned DB2 UNIX databases.7 DB2/UDB Enterprise stage options The DB2/UDB Enterprise (native parallel) stage should be used for reading from. or translated into any language in any form by any means without the written permission of IBM. 2006 108 of 179 © 2006 IBM Information Integration Solutions. and writing to a DB2 Enterprise Server Edition database with Database Partitioning Feature (DPF) • As a native. The DB2 load operator performs a non-recoverable load. All activity in the z/OS environment always goes through the DB2 coordinator node so parallelism differs slightly depending on how DB2 is accessed. if the load operation is terminated before it is completed. 10. performing lookups against.9 DB2 in the DataStage USS environment The manner in which DataStage / USS Edition interfaces with DB2 is slightly different than it is in the non-z/OS environment.2. No part of this publication may be reproduced.2. . reading from and writing to DB2 in parallel (where appropriate). parallel component.Information Integration Solutions Center of Excellence 10. In this scenario. it may be beneficial to have the DB2 DBA configure separate DB2 coordinator nodes (no local data) on each ETL server (in clustered ETL configurations). • DB2/UDB Enterprise stage is tightly integrated with the DB2 RDBMS. a) The Write Method (and corresponding insert / update / upsert / delete) communicates directly with the DB2 database nodes to execute instructions in parallel.8 Performance Notes In some cases. All operations are logged to the DB2 database log. and using the same data partitioning as the referenced DB2 tables. b) The DB2 Load method requires that the DataStage user running the job have DBADM privilege on the target DB2 database. When writing to a DB2 database in parallel.

functions within the db2read operator are used to read the DB2 SYSTABLES table to retrieve the tablespace and database name for the table. and is controlled by the number of nodes in the configuration file. Parallel Framework Red Book: Data Flow Job Design July 17. This is illustrated in Figure 47. transmitted. Since all write operations need to go through the DB2 coordinator node on z/OS (this is different than on non-z/OS platforms). stored in a retrieval system. Table T is in tablespace TS and TS is partitioned into 3 partitions on Col1 (limits: F. 40). P. as illustrated in Figure 46: Figure 46: DB2 read on DataStage/USS For example. All rights reserved. 30. partitioning index name(s). This information determines the number of db2read operators that the conductor builds into the score and the queries that they execute. T) and Col2 (10. 2006 109 of 179 © 2006 IBM Information Integration Solutions. . These values are in turn used to read the SYSTABLEPART table to retrieve the number of partitions. the number of operators do not have to match to the number of partitions. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. transcribed. The WHERE clauses which are created to read this tables are: Where Col1 < ‘F’ or (Col1 = ‘F’ and (Col2 < 10 or Col2 = 10)) Where (Col1 > ‘F’ and Col1 < ‘P’) or (Col1 = ‘F’ and Col2 > 10) or (Col1 = ‘P’ and (Col2 < 20 or Col2 = 20)) Where Col1 > ‘T’ or (Col1 = ‘T’ and Col2 > 40) The method that DataStage/USS Edition uses to write to DB2 UDB on z/OS works differently than the read process. Finally the SYSKEYS and SYSCOLUMNS tables are read using the index name to get the associated column metadata (name and type). and the partition limit key value(s).Information Integration Solutions Center of Excellence When accessing a DB2 table using the Table read method.

2006 110 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. . transmitted. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence Figure 47: DB2 write on DataStage/USS On DataStage/USS Edition. An example of an in-memory Normal Lookup is shown in Figure 48. No part of this publication may be reproduced. Lookups work differently depending on whether the lookup is done normally (in memory) or using a sparse technique where each lookup is effectively a query to the database. All rights reserved. transcribed.

No part of this publication may be reproduced. The LOAD utility has a second limitation in that data cannot be piped into it. 2006 111 of 179 © 2006 IBM Information Integration Solutions. Contrast the Normal Lookup with the way a Sparse Lookup is done as shown in Figure 49.Information Integration Solutions Center of Excellence Figure 48: In-Memory Lookup on DataStage/USS Here we see that the Normal Lookup actually consists of reading the DB2 table into memory and then performing the lookup against the memory copy of the table. In order to invoke it from a DataStage/USS job. stored in a retrieval system. nor can it be read in from a USS HFS file. All rights reserved. The DB2 LOAD utility is designed to run from JCL only. Figure 50 illustrates the DB2 LOAD process on USS and also shows the format of the special resource statement used to define the MVS dataset used during the load operation. Figure 49: DB2 Sparse Lookup on DataStage/USS Finally. using the DB2 load utility in USS is different from non-z/OS environments. where each lookup operator is issuing an SQL to DB2 for every row it processes. . Since each of these queries must go through the DB2 coordinator node we can effectively ignore the level of parallelism specified for the table. This requires DataStage/USS to create an MVS flat file to pass to the loader – note that this is the only non-HFS file that DS/USS can write to. Since there is no sequential file stage associated with this MVS load file. we need to add a special resource statement in our configuration file to specify the MVS dataset name to use. it matches the number of db2read operators to the partitioning scheme of the table (similar to the read) and the number of lookup operators to the number of nodes in the configuration file. we call a DB2 stored procedure called DSNUTILS. Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. transcribed. When the conductor creates the score. transmitted.

Information Integration Solutions Center of Excellence Figure 50: Calling DB2 Load Utility on DataStage/USS Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. 2006 112 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. . All rights reserved. transmitted. transcribed.

If the DATETIME starts with a year component.3.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: .columns of the database row correspond to columns of a DS/EE record . and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. DECIMAL[p. or translated into any language in any form by any means without the written permission of IBM.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n.rows of the database result set correspond to records of a DS/EE Data Set . decimal[p. time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month.3. No part of this publication may be reproduced. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types . Informix Data Type CHAR(n) CHARACTER VARYING(n.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. as shown in the following table.r) NVARCHAR(n.3 Informix Database Guidelines 10. .r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date. The Informix Enterprise stage converts Informix data types to Enterprise Edition data types.Information Integration Solutions Center of Excellence 10.both DS/EE and Informix support null values. the result is a timestamp field. stored in a retrieval system. transmitted. the result is a date field. 2006 113 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. the result is a time field. transcribed. A decimal with floating scale is converted to dfloat dfloat dfloat int32 decimal string[n] string[max=n] sfloat int32 sfloat int16 string[max=n] IMPORTANT: Informix data types that are not listed in the above table cannot be used in the Informix Enterprise stage.2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.s] where p is the precision and s is the scale The maximum precision is 32.

transcribed.2 ODBC Enterprise stage Data Type Mapping ODBC data sources are not supported by the orcdbutil utility.s] decimal[p.4. In that case. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.s] string[36] int32 int8 [0 or 1] decimal[p. two underscore characters replace the unsupported character . . No part of this publication may be reproduced. as shown in the following table: ODBC Data Type SQL_BIGINT SQL_BINARY SQL_CHAR SQL_DECIMAL SQL_DOUBLE SQL_FLOAT SQL_GUID SQL_INTEGER SQL_BIT SQL_REAL SQL_SMALLINT SQL_TINYINT SQL_TYPE_DATE SQL_TYPE_TIME SQL_TYPE_TIMESTAMP SQL_VARBINARY SQL_VARCHAR SQL_WCHAR SQL_WVARCHAR Enterprise Edition Data Type int64 raw(n) string[n] decimal[p.s] where p is the precision and s is the scale decimal[p.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types . Enterprise Edition will return an error and abort your job. or translated into any language in any form by any means without the written permission of IBM. transmitted.4.s] int16 int8 date time[p] timestamp[p] raw[max=n] string[max=n] ustring[n] ustring[max=n] Note that the maximum size of a DataStage record is limited to 32K. If you attempt to read a record larger than 32K. It is important to verify the correct ODBC to Enterprise Edition data mapping.names are translated exactly except when the external data source column name contains a character that DataStage does not support. stored in a retrieval system.rows of the database result set correspond to records of a DS/EE Data Set .columns of the database row correspond to columns of a DS/EE record .Information Integration Solutions Center of Excellence 10.both DS/EE and ODBC support null values.4 ODBC Enterprise Guidelines 10. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17.1 ODBC Enterprise Stage Column Names For each ODBC Enterprise stage: . 2006 114 of 179 © 2006 IBM Information Integration Solutions.

Information Integration Solutions Center of Excellence IMPORTANT: ODBC data types that are not listed in the above table cannot be used in the ODBC Enterprise stage. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. transmitted. . stored in a retrieval system. 2006 115 of 179 © 2006 IBM Information Integration Solutions. transcribed. All rights reserved.

the ODBC Enterprise stage does not support parallel read4. transcribed. . Depending on the target database. since this capability is not provided by the ODBC API. or translated into any language in any form by any means without the written permission of IBM. transmitted. All rights reserved. it may be possible to write to a target database in parallel using the ODBC Enterprise stage.Information Integration Solutions Center of Excellence 10. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17.4. a patch may be available through IBM IIS Support to support parallel reads through ODBC. and the table configuration (row or page level lock mode if available). 4 On some platforms. Parallel reads through ODBC match the degree of parallelism in the $APT_CONFIG_FILE. stored in a retrieval system. 2006 116 of 179 © 2006 IBM Information Integration Solutions.3 Reading ODBC Sources Unlike other native parallel database stages.

transmitted.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p. Examples include: . The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition.names are translated exactly except when the Oracle source column name contains a character that DataStage does not support.s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38.5.3 Reading from Oracle in Parallel By default. transcribed.both DS/EE and Oracle support null values. which are not heavily typed.5 Oracle Database Guidelines 10.queries containing a GROUP BY clause that are also hash partitioned on the same field Parallel Framework Red Book: Data Flow Job Design July 17. 2006 117 of 179 © 2006 IBM Information Integration Solutions. .2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.Information Integration Solutions Center of Excellence 10.5.5. stored in a retrieval system. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. Enterprise Edition will return an error and abort your job. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage. the Oracle Enterprise stage reads sequentially from its source table or query.columns of the database row correspond to columns of a DS/EE record . and will generate an error at runtime 10. two underscore characters replace the unsupported character .s] if precision (p) >=11 or scale > 0 not supported string[max=n] or ustring[max=n] a variable-length string with maximum length = n Note that the maximum size of a DataStage record is limited to 32K. It is important to note that certain types of queries cannot run in parallel. This is particularly important for Oracle databases. All rights reserved. Setting the partition table option to the specified table will enable parallel extracts from an Oracle source. Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p.1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . or translated into any language in any form by any means without the written permission of IBM.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types . No part of this publication may be reproduced. If you attempt to read a record larger than 32K.rows of the database result set correspond to records of a DS/EE Data Set . In that case.

Information Integration Solutions Center of Excellence - queries performing a non-collocated join (a SQL JOIN between two tables that are not stored in the same partitions with the same partitioning strategy) 10. 2006 118 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. rebuild). The Upsert Write Method can be used to insert rows into a target Oracle table without bypassing indexes or constraints. Enterprise Edition uses the Parallel Direct Path Load method. . transcribed. In this instance. the Oracle load will be done sequentially. PARALLEL=FALSE) also allows loading of indexed tables without index maintenance. All rights reserved. In order to automatically generate the SQL required by the Upsert method. the key column(s) must be identified using the check boxes in the column grid. or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. • Setting the environment variable $APT_ORACLE_LOAD_OPTIONS to “OPTIONS (DIRECT=TRUE.4 Oracle Load Options When writing to an Oracle table (using Write Method = Load). stored in a retrieval system. transmitted. When using this method.5. the Oracle stage cannot write to a table that has indexes (including indexes automatically generated by Primary Key constraints) on it unless you specify the Index Mode option (maintenance.

Information Integration Solutions Center of Excellence

10.6 Sybase Enterprise Guidelines
10.6.1 Sybase Enterprise Stage Column Names For each Sybase Enterprise stage: - rows of the database result set correspond to records of a DS/EE Data Set - columns of the database row correspond to columns of a DS/EE record - the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types - names are translated exactly except when the Sybase source column name contains a character that DataStage does not support. In that case, two underscore characters replace the unsupported character - both DS/EE and Sybase support null values, and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.6.2 Sybase Enterprise stage Data Type Mapping Sybase databases are not supported by the orcdbutil utility. It is important to verify the correct Sybase to Enterprise Edition data mapping, as shown in the following table: Sybase Data Type
BINARY(n) BIT CHAR(n) DATE DATETIME DEC[p,s] or DECIMAL[p,s] DOUBLE PRECISION or FLOAT INT or INTEGER MONEY NCHAR(n) NUMERIC[p,s] NVARCHAR(n,r) REAL SERIAL SMALLDATETIME SMALLFLOAT SMALLINT SMALLMONEY TINYINT TIME UNSIGNED INT VARBINARY(n) VARCHAR(n)

Enterprise Edition Data Type
raw(n) int8 string[n] a fixed-length string with length n date timestamp decimal[p,s] where p is the precision and s is the scale dfloat int32 decimal[15,4] ustring[n] a fixed-length string with length n - only for ASE decimal[p,s] where p is the precision and s is the scale ustring[max=n] a variable-length string with length n - only for ASE sfloat int32 timestamp sfloat int16 decimal[10,4] int8 time unit32 raw[max=n] string[max=n] a variable-length string with maximum length n

IMPORTANT: Sybase data types that are not listed in the above table cannot be used in the Sybase Enterprise stage, and will generate an error at runtime
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 119 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

10.7 Teradata Database Guidelines
10.7.1 Choosing the Proper Teradata Stage Within DataStage Enterprise Edition, the following stages can be used for reading from and writing to Teradata databases in a parallel job flow: Source Teradata Stages Teradata Enterprise Teradata API Target Teradata Stages Teradata Enterprise Teradata API Teradata MultiLoad (MultiLoad option) Teradata MultiLoad (TPump option)

For maximum performance of high-volume data flows, the native parallel Teradata Enterprise stage should be used. Teradata Enterprise uses the programming interface of the Teradata utilities FastExport (reads) and FastLoad (writes), and is subject to all these utilities’ restrictions. NOTE: Unlike the FastLoad utility, the Teradata Enterprise stage supports Append mode, inserting rows into an existing target table. This is done through a shadow “terasync” table.

Teradata has a system-wide limit to the number of concurrent database utilities. Each use of the Teradata Enterprise stages counts toward this limit. 10.7.2 Source Teradata Stages Teradata Stage Stage Usage Guidelines Type
Teradata Enterprise Teradata API Native Parallel Plug-In - Reading a large number of rows in parallel - Supports OPEN and CLOSE commands - Subject to the limits of Teradata FastExport - Reading a small number of rows sequentially

Parallel Read
Yes No

Teradata Utility Limit
applies none

10.7.3 Target Teradata Stages Teradata Stage Stage Usage Guidelines Type
Teradata Enterprise Native Parallel - Writing a large number of rows in parallel - Supports OPEN and CLOSE commands - Limited to INSERT (new table) or APPEND (existing table) - Subject to the limits of Teradata FastLoad (but also supports APPEND) - Locks the target table in exclusive mode - Insert, Update, Delete, Upsert of moderate data volumes - Locks the target table(s) in exclusive mode

Parallel Write
Yes

Teradata Utility Limit
applies

Teradata MultiLoad (MultiLoad utility)

Plug-In

No

applies

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

120 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence Teradata MultiLoad (TPump utility) Plug-In - Insert, Update, Delete, Upsert of small volumes of data within a large database - Does not lock the target tables - Should not be run in parallel, because each node and use counts toward system-wide Teradata utility limit - Insert, Update, Delete, Upsert of small volumes of data - Allows concurrent writes (does not lock target) - Slower than TPump for equivalent operations

Teradata API

Plug-In

Yes

none

10.7.4 Teradata Enterprise Stage Column Names For each Teradata Enterprise stage: - rows of the database result set correspond to records of a DS/EE Data Set - columns of the database row correspond to columns of a DS/EE record - the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types - both DS/EE and Teradata support null values, and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column - DS/EE gives the same name to its columns as the Teradata column name. However, while DS/EE column names can appear in either upper or lower case, Teradata column names appear only in upper case. 10.7.5 Teradata Enterprise stage Data Type Mapping Teradata databases are not supported by the orcdbutil utility. It is important to verify the correct Teradata to Enterprise Edition data mapping, as shown in the following table: Teradata Data Type
byte(n) byteint char(n) date decimal[p,s] double precision float graphic(n) integer long varchar long vargraphic numeric(p,s) real smallint time timestamp varbyte(n) varchar(n)

Enterprise Edition Data Type
raw[n] int8 string[n] date decimal[p,s] where p is the precision and s is the scale dfloat dfloat raw[max=n] int32 string[max=n] raw[max=n] decimal[p,s] Dfloat int16 time timestamp raw[max=n] string[max=n] July 17, 2006 121 of 179

Parallel Framework Red Book: Data Flow Job Design

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

SessionsPerPlayer=nn][. The SessionsPerPlayer option determines the number of connections each DataStage EE player opens to Teradata. transmitted.Information Integration Solutions Center of Excellence vargraphic(n) raw[max=n] IMPORTANT: Teradata data types that are not listed in the above table cannot be used in the Teradata Enterprise stage. 10. By default. and/or RequestedSessions should be decreased. RequestedSessions equals the maximum number of available sessions on the Teradata instance. In that case SessionsPerPlayer should be increased. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. transcribed. Aggregates and most arithmetic operators are not allowed in the SELECT clause of a Teradata Enterprise stage. No part of this publication may be reproduced. Teradata Enterprise stage uses 32K buffers. .6 Specifying Teradata Passwords with Special Characters Teradata permits passwords with special characters and symbols. this determines the number of DataStage players. and hence the number of UNIX processes and overall system resource requirements of the DataStage job. but this can be set to a value between 1 and the database vprocs. and will generate an error at runtime. Indirectly. SessionsPerPlayer should be set such that: RequestedSessions = (sessions per player * the number of nodes * players per node) The default value for the SessionsPerPlayer suboption is 2.7 Teradata Enterprise Settings Within the Teradata Enterprise stage. To specify a Teradata password that contains special characters.password=password[. (Note that 64K buffers must be enabled at the Teradata server level). Parallel Framework Red Book: Data Flow Job Design July 17.RequestedSessions=nn] where SesionsPerPlayer and RequestedSessions are optional connection parameters that are required when accessing large Teradata databases.7. 10. the password must be surrounded by an “escaped” single quote as shown. stored in a retrieval system. the DB Options property specifies the connection string and connection properties in the form: user=username. 2006 122 of 179 © 2006 IBM Information Integration Solutions. By default. where pa$$ is the example password: \’pa$$\’ 10.7.7. Setting the SessionsPerPlayer too low on a large system can result in so many players that the job fails due to insufficient resources.8 Improving Teradata Enterprise Performance Setting the environment variable $APT_TERA_64K_BUFFERS may significantly improve performance of Teradata Enterprise connections depending on network configuration.

To connect to a Teradata server. where x is 0-9. also known as the tdpid. you must supply the client with the Teradata Director Program (TDP) identifier. That leaves 39 possible TDP names and is different than the convention used for non-channel attached systems. stored in a retrieval system. No part of this publication may be reproduced. the tdpid must be in the form TDPx. The first three characters must be TDP. the tdpid is the host name of the Teradata server. transcribed. 2006 123 of 179 © 2006 IBM Information Integration Solutions.9 Teradata on USS On the USS platform the Teradata Enterprise Stage uses CLIv2 for channel-attached systems (OS/390 and z/OS). Parallel Framework Red Book: Data Flow Job Design July 17. #. A-Z (case insensitive). $. or translated into any language in any form by any means without the written permission of IBM. . or @.7. On a network-attached system.Information Integration Solutions Center of Excellence 10. transmitted. All rights reserved. On MVS.

the job’s partitioning logic and parallel design concepts (especially within Transformer stages) should be examined. transcribed. July 17. Using the job monitor performance statistics. Prints detailed information in the job log for each operator when allocating additional heap memory. it is important that all jobs be tested with a configuration file that has more than one node in its default node pool. including CPU utilization and elapsed processing time. When trying to start very large jobs on heavilyloaded servers. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 11 Troubleshooting and Monitoring 11. or translated into any language in any form by any means without the written permission of IBM. and internal Enterprise Edition log messages in a directory corresponding to the job name. this can identify which part of a job flow is impacting overall performance. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. $DS_PX_DEBUG 1 $APT_PM_STARTUP_CONCURRENCY 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Normally. 2006 124 of 179 $APT_PM_PLAYER_TIMING 1 1 FORCE $APT_PM_PLAYER_MEMORY $APT_BUFFERING_POLICY Setting $APT_BUFFERING_POLICY=FORCE is not recommended for production job runs. If the job results are correct with a single-node configuration file. Set this environment variable to capture copies of the job score. generated osh. the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance.2 Debugging Environment Variables The following environment variables can be set to assist in debugging a parallel job: Environment Variable $OSH_PRINT_SCHEMAS Setting 1 Description Outputs the actual schema definitions used by the DataStage EE framework at runtime in the DataStage log.1 Warning on Single-Node Configuration Files Because the DS/EE configuration file can be changed at runtime. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. stored in a retrieval system. This will ensure that the jobs have been designed with proper partitioning logic. Forces an internal buffer operator to be placed between every operator. Prints detailed information in the job log for each operator. 11. transmitted. This can be useful when determining if the actual runtime schema matches the expected job design table definitions. and incorrect with a multi-node configuration file. lowering this number will limit the number of processes that are simultaneously created when a job is started. This environment variable should not normally need to be set. All rights reserved. .

it is possible for $OSH_PRINT_SCHEMAS to generate a log entry that is too large for DataStage Director to store or display. In some instances. 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. and internal Enterprise Edition log messages in a directory corresponding to the job name. These may indicate an underlying logic problem or unexpected data type conversion. it may not be possible to remove all warning messages generated by the EE engine. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. stored in a retrieval system. The general process for debugging a job is:  Check the Director job log for warnings. look at row counts.  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled). The default is 30 seconds. transcribed. transmitted. To capture the full schema output in these cases. the log entry is sometimes preceded by a warning condition. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. Use the Data Set Management tool (available in the Tools menu of DataStage Designer or DataStage Manager) to examine the schema. This will place entries in the Director log with the actual runtime schema for every link using Enterprise Edition internal data types. and to manage source or target Parallel Data Sets. or translated into any language in any form by any means without the written permission of IBM. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. All rights reserved.5. All fatal and warning messages should be addressed before attempting to debug. 11. But all warnings should be examined and understood.Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. tune.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide. . generated osh. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG .  Enable the Job Monitoring Environment Variables detailed in Section 2.3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. When a fatal error occurs. or promote a job from development into test or production. No part of this publication may be reproduced. Set the environment variable $DS_PX_DEBUG to capture copies of the job score. July 17.    NOTE: For large jobs.

To enable viewing of generated OSH. It is useful to examine the generated OSH for debugging purposes. NOTE: The wc command counts UNIX line delimiters. it must be enabled for a given project within the Administrator client: Figure 51: Generated OSH Administrator option Once this option has been enabled for a given project. this count may be incorrect. transmitted. . It is also not useful for files of non-delimited fixed-length record format. transcribed. jobs are compiled into OSH (Orchestrate SHell) scripts that are used to execute the given job design at runtime. 2006 126 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. All rights reserved.Information Integration Solutions Center of Excellence  For flat (sequential) sources and targets: o To display the actual contents of any file in hexadecimal and ASCII (including embedded control characters or ASCII NULLs). No part of this publication may be reproduced.4 Viewing the Generated OSH Within Designer. stored in a retrieval system. so if the file has any binary columns. use the UNIX command wc –lc [filename] Dividing the total number of characters number of lines provides an audit to ensure all rows are same length. use the UNIX command od –xc –Ax o To display the number of lines and characters in a specified ASCII text file. 11. the generated OSH tab will appear in the Job Properties dialog box: Parallel Framework Red Book: Data Flow Job Design July 17. and to understand internally what is running.

All rights reserved. have been inserted by EE. that is. transmitted. The Peek and Sort stages are combined. All stages in this flow are running on one node. the first task is to examine the score dump which is generated when you set APT_DUMP_SCORE=TRUE in your environment. transcribed. It shows three stages: Generator. 2006 127 of 179 © 2006 IBM Information Integration Solutions. where data is repartitioned and how it is repartitioned. A score dump includes a variety of information about a flow. The job runs 3 processes on 2 nodes. including how composite operators and shared containers break down. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Figure 52: Generated OSH in Designer Job Properties 11. and exactly which nodes each operator runs on.5 Interpreting the Parallel Job Score When attempting to understand an Enterprise Edition flow. which operators. ##I TFSC 004000 14:51:50(000) <main_program> This step has 1 dataset: ds0: {op0[1p] (sequential generator) eOther(APT_HashPartitioner { key={ value=a } })->eCollectAny op1[2p] (parallel APT_CombinedOperatorController:tsort)} It has 2 operators: op0[1p] {(sequential generator) on nodes ( Parallel Framework Red Book: Data Flow Job Design July 17. . stored in a retrieval system. what degree of parallelism each operator runs with. Also available is some information about where data may be buffered. The following score dump shows a flow with a single Data Set. if any. they have been optimized into the same process. No part of this publication may be reproduced. which has a hash partitioner that partitions on key field a. Sort (tsort) and Peek.

there are three areas to investigate: • Are there sequential stages? • Is needless repartitioning occurring? • In a cluster or Grid.p0] )} op1[2p] {(parallel APT_CombinedOperatorController: (tsort) (peek) )on nodes ( lemond.p0] lemond.4.torrent.2 Understanding the Parallel Job Score.torrent. transcribed. No part of this publication may be reproduced. transmitted. are the computation-intensive stages shared evenly across all nodes? More details on interpreting the parallel job score can be found in 12.p1] )} In a score dump.torrent.Information Integration Solutions Center of Excellence lemond. stored in a retrieval system.com[op0. All rights reserved.com[op1. Parallel Framework Red Book: Data Flow Job Design July 17. . 2006 128 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM.com[op1.

No part of this publication may be reproduced.1 How to Design a Job for Optimal Performance Overall job design can be the most significant factor in data flow performance. 12. use a select list to read needed columns instead of the entire table (if possible) • Be alert when using runtime column propagation (“RCP”) – it may be necessary to disable RCP for a particular stage to ensure that columns are actually removed using that stage’s Output Mapping. This section outlines performance-related tips that can be followed when building a parallel data flow using DataStage Enterprise Edition. stored in a retrieval system. a) Use Parallel Data Sets to land intermediate result between parallel jobs. . • Lookup File Sets can be used to store reference data used in subsequent jobs. 2006 129 of 179 © 2006 IBM Information Integration Solutions. and for optimizing the performance of a given data flow using various settings and features within DataStage Enterprise Edition. Every unused column requires additional memory which can impact performance (it also makes each transfer of a record from one stage to the next more expensive). facilitating end-to-end parallelism across job boundaries. • Parallel Data Sets retain data partitioning and sort order.Information Integration Solutions Center of Excellence 12 Performance Tuning Job Designs The ability to process large volumes of data in a short period of time depends on all aspects of the flow and environment being optimized for maximum throughput and performance. Lookup File Sets can only be used on reference links to a Lookup stage. • When reading from database sources. This section provides tips for designing a job for optimal performance. They maintain reference data in DS/EE internal format. b) Remove unneeded columns as early as possible within the data flow. Parallel Framework Red Book: Data Flow Job Design July 17. Performance tuning and optimization is an iterative process that begins at job design and unit tests. proceeds through integration and volume testing. However. File Sets facilitate parallel I/O at the expense of exporting to a specified file format. If you need to share information with external applications. and continues throughout an application’s production lifecycle. or translated into any language in any form by any means without the written permission of IBM. There are no utilities for examining data within a Lookup File Set. • Data Sets can only be read by other DS/EE parallel jobs (or the orchadmin command line utility). pre-indexed. transmitted. transcribed. All rights reserved. in the DS/EE native internal format.

 Enable $OSH_PRINT_SCHEMAS to verify runtime schema matches job design column definitions o Verify that the data type of defined Transformer stage variables matches the expected result type e) Minimize the number of Transformers. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. stored in a retrieval system. Because BuildOps are built in C++. that unless dynamic (parameterized) conditions are required. especially in large-volume data flows. a Transformer is always faster than a Filter or Switch stage. if possible. It is usually possible to choose a smaller partition-key set. 2006 130 of 179 © 2006 IBM Information Integration Solutions. External user-defined functions can expand the capabilities of the parallel Transformer. o When working with database sources and targets. . and to simply re-sort on a differing set of secondary/tertiary/etc. transmitted. When business rules dictate otherwise and the data volume is large and sufficiently skewed. renaming and removing columns. keys. Use BuildOps only when existing Transformers do not meet performance requirements. For example:  Varchar columns of a large (for example. For data type conversions. other stages (for example. • • Avoid using the BASIC Transformer. however. transcribed. • There are limited scenarios when the memory overhead of handling large Varchar columns would dictate the use of unbounded strings. f) Minimize the number of partitioners in a job. use orchdbutil to ensure that the designtime metadata matches the actual runtime metadata (especially with Oracle databases). Modify) may be more appropriate. there is greater control over the efficiency of code. 32K) maximum length that are rarely populated  Varchar columns of a large maximum length with highly varying data sizes d) Avoid type conversions. Copy. Unbounded strings (Varchar’s without a maximum length) can have a significant negative performance impact on a job flow. at the expense of ease of development (and more skilled developer requirements). repartition to a more balanced distribution as soon as possible to improve performance of downstream stages. ensure data is as close to evenly distributed as possible. or translated into any language in any form by any means without the written permission of IBM. • When possible. or when complex reusable logic is required. No part of this publication may be reproduced. Note.Information Integration Solutions Center of Excellence c) Always specify a maximum length for Varchar columns.

Parallel Framework Red Book: Data Flow Job Design July 17. not explicitly stated in the sort keys) row order. Remember that SAME maintains the degree of parallelism of the upstream operator. o Performance of individual sorts can be improved by increasing the memory usage per partition using the “Restrict Memory Usage (MB)” option of the standalone Sort stage. and groupings. disabling operator combination allows CPU activity to be spread across multiple processors instead of being constrained to a single processor. transcribed. previously sorted” option for those key columns in the Sort stage will reduce the cost of sorting and take greater advantage of pipeline parallelism. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. If two processes are interdependent (one processes the other’s output) and they are both CPU-bound or I/O-bound.Information Integration Solutions Center of Excellence • • Know your data. Choose hash key columns that generate sufficient unique key combinations (while satisfying business requirements). Use SAME partitioning carefully. . and should only be used if there is a need to maintain an implied (i. specifying the “don’t sort. o The stable sort option is much more expensive than non-stable sorts. As with any example. In addition. if possible.  If data has already been partitioned and sorted on a set of key columns. o When writing to parallel Data Sets. by using “SAME” partitioning. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be used to set (in units of MB) the size of the RAM buffer for all sorts in a job. transmitted. Combined operators are generally adjacent to each other in a data flow. even those that have the “Restrict Memory Usage” option set. partitioning. your results should be tested in your environment. there is nothing to be gained from pipeline partitioning5. When composing the score. 12. In these configurations. The purpose behind operator combination is to reduce the overhead associated with an increased process count. The default setting is 20MB per partition. stored in a retrieval system. try to maintain this sorting. g) Minimize and combine use of Sorts where possible o It is frequently possible to arrange the order of business logic within a job flow to leverage the same sort order. Enterprise Edition attempts to reduce the number of processes by combining the logic from 2 or more stages (operators) into a single process (per partition). DataStage Enterprise Edition analyzes a given job design and uses the parallel configuration file to build a job score which defines the processes and connection topology (Data Sets) between them used to execute the job logic. 5 One exception to this guideline is when operator combination generates too few processes to keep the processors busy. sort order and partitioning are preserved. All rights reserved.e. 2006 131 of 179 © 2006 IBM Information Integration Solutions. When reading from these Data Sets.2 Understanding Operator Combination At runtime.

transmitted. stored in a retrieval system. as shown in the following illustration: Figure 53: CPU-bound combined process in Job Monitor Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM.) In addition. this is only recommended on pre 7. see 12. the assumptions used by the Enterprise Edition optimizer to determine which stages can be combined may not always be the most efficient.2 Understanding the Parallel Job Score” in this document. Combined Operator Controller). combined stages are indicated by parenthesis surrounding the % CPU. . All rights reserved.0 versions of DS/EE. or globally. 2006 132 of 179 © 2006 IBM Information Integration Solutions. if the “%CPU” column is displayed in a Job Monitor window in Director. (For information on interpreting a job score dump. disables ALL combination in the entire data flow. when other performance tuning measures have been applied and still greater performance is needed. here is a partial list of non-combinable operators:  Join  Aggregator  Remove Duplicates  Merge  BufferOp  Funnel  DB2 Enterprise Stage  Oracle Enterprise Stage  ODBC Enterprise Stage  BuildOps In general.4.a.Information Integration Solutions Center of Excellence However.it will include all operators that meet the following rules: o Must be contiguous o Must be the same degree of parallelism o Must be ‘Combinable’. It is for this reason that combination can be enabled or disabled on a per-stage basis. Enterprise Edition is ‘greedy’ . No part of this publication may be reproduced. However. transcribed. combination can be set on a per-stage basis (on the Stage/Advanced tab) The job score identifies what components are combined.k. There are 2 ways to affect operator combination: o The environment variable APT_DISABLE_COMBINATION. When deciding which operators to include in a particular combined operator (a. tuning combination might yield additional performance benefits. o Within Designer. it is best to let DSEE decide what to combine and what to leave uncombined.

transcribed. 2006 133 of 179 © 2006 IBM Information Integration Solutions. it might be a good idea to set the sequential file to be non-combinable. Toward that end. There are many factors that can reduce the number of processes generated at runtime:  Use a single-node configuration file  Remove ALL partitioners and collectors (especially when using a single-node configuration file)  Enable runtime column propagation on Copy stages with only one input and one output  Minimize join structures (any stage with more than one input. such as Join. In fact. it is good to separate I/O heavy operators (Sequential File. DS/EE executes a given job across the resources defined in a the specified configuration file. Merge. Disabling combination with the Transformer enables pipeline partitioning. However.Information Integration Solutions Center of Excellence Choosing which operators to disallow combination for is as much art as science. . Full Sorts. in general.) from CPU-heavy operators (Transformer. etc.). when it is appropriate to minimize the resource requirements for a given scenario. This will prevent IO requests from waiting on CPU to become available and viceversa. in the above job design. for example:  Batch jobs that process a small volume of data  Real-time jobs that process data in small message units  Environments running a large number of jobs simultaneously on the same server(s) In these instances. No part of this publication may be reproduced.3 Minimizing Runtime Processes and Resource Requirements The architecture of Enterprise Edition is well suited for processing massive volumes of data in parallel across available resources. Funnel) Parallel Framework Red Book: Data Flow Job Design July 17. There are times. as shown in this subsequent Job Monitor for the same job: Figure 54: Throughput in Job Monitor after disabling combination 12. etc. or translated into any language in any form by any means without the written permission of IBM. and improves performance. Change Capture. a single-node configuration file is often appropriate to minimize job startup time and resource requirements without significantly impacting overall performance. All rights reserved. For example. stored in a retrieval system. Lookup. however. if you have several transformers and database operators combined with an output Sequential File. the I/O-intensive FileSet is combined with a CPU-intensive Transformer. transmitted.

strictly speaking. Once the upstream block is full and the downstream block is empty.4. transmitted. All rights reserved. . Each pair of operators that have a producer/consumer relationship will share at least 2 of these blocks. or translated into any language in any form by any means without the written permission of IBM. The second block will be used by the downstream/consumer stage to obtain data that is ready for the next processing step. transcribed. so setting this value too high will result in a large amount of memory consumption. DB2 Enterprise. Oracle Enterprise. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MAX_TRANSPORT_BLOCK_SIZE APT_MAX_TRANSPORT_BLOCK_SIZE July 17. APT_MIN_TRANSPORT_BLOCK_SIZE o Specifies the minimum allowable block size for transferring data between players. 2006 134 of 179   Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. ODBC Enterprise. 12. then setting APT_DEFAULT_TRANSPORT_BLOCK_SIZE to a multiple of (or equal to) the record size will resolve the problem. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. the value provided by a user for this variable is rounded up to the operating system's nearest page size.1 Inter-Operator Transport Buffering Though it may look like it from the performance statistics and documentation might discuss ‘record streaming’. Remember. The behavior of these transport blocks is determined by these environment variables:  APT_DEFAULT_TRANSPORT_BLOCK_SIZE o Specifies the default block size for transferring data between players. If necessary. (Buffering is discussed in more detail in the following section. Merge.) 12. the blocks are swapped and the process begins again. Default is 8192. The first block will be used by the upstream/producer stage to output data it is done with. Remove Duplicates. there are 2 of these transport blocks for each partition of each link. This type of buffering (or ‘Record Blocking’) is rarely tuned.Information Integration Solutions Center of Excellence   Minimize non-combinable stages (as outlined in the previous section) such as Join. They are actually transferred in blocks (just like on old magnetic tapes) called “Transport Blocks”. stored in a retrieval system. Aggregator. BuildOps. No part of this publication may be reproduced. records do not stream from one stage to another. Cannot be less than 8192. BufferOp Selectively (being careful to to avoid deadlocks) disable buffering. with a valid value range for between 8192 and 1048576.4 Understanding Buffering There are two types of buffering in Enterprise Edition: ‘inter-operator transport’ and ‘deadlock prevention’. The default value is 8192. It usually only comes into play when the size of a single record exceeds the default size of the transport block. cannot be greater than 1048576. Funnel.

the framework calculates the block size for transferring data between players according to this algorithm: if (recordSize * APT_LATENCY_COEFFICIENT < APT_MIN_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MIN_TRANSPORT_BLOCK_SIZE else if (recordSize * APT_LATENCY_COEFFICIENT > APT_MAX_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MAX_TRANSPORT_BLOCK_SIZE else blockSize = recordSize * APT_LATENCY_COEFFICIENT APT_LATENCY_COEFFICIENT o Specifies the number of records to be written to each transport block APT_SHARED_MEMORY_BUFFERS o Specifies the number of Transport Blocks between a pair of operators. cannot be greater than 1048576. transmitted. APT_LATENCY_COEFFICIENT. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE. stored in a retrieval system. No part of this publication may be reproduced. Default is 1048576. All rights reserved. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records. . 12. “Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job.4. 2006 135 of 179 © 2006 IBM Information Integration Solutions. transcribed.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players. Cannot be less than 8192. or translated into any language in any form by any means without the written permission of IBM. Here is an example job fragment: Figure 55: Fork-Join example Parallel Framework Red Book: Data Flow Job Design July 17.2 Deadlock Prevention Buffering The other type of buffering.

To understand deadlock-prevention buffering.) Without deadlock buffering. and Join is waiting to read from Aggregator2. Aggregator2 is waiting to read from the Transformer. it would eventually time out). To guarantee that this problem never happens in Enterprise Edition. or translated into any language in any form by any means without the written permission of IBM. BufferOp is always ready to read or write and will not allow a read/write request to be queued. transcribed. it is important to understand the operation of a parallel pipeline. which go into an Inner Join. stored in a retrieval system. Note however. the Transformer creates a fork with 2 parallel Aggregators. instead of data flow. Without deadlock buffering. d eue Qu ad Re Transformer Waiting to write to Aggregator1 Qu eue Re d ad Join Waiting to read from Aggregator2 July 17. Aggregator1 is waiting on Join. there is a specialized operator called BufferOp.bringing processing to a halt (though the job does not stop running. No part of this publication may be reproduced. So the above job structure would be altered by the DS/EE engine to look like this: Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. All rights reserved. 2006 136 of 179 . Imagine that the Transformer is waiting to write to Aggregator1. transmitted. Join is waiting on Aggregator2.Information Integration Solutions Center of Excellence In this example. Like this: Aggregator 1 Waiting to Write to Join d e ue Qu ite Wr Qu eu Wr ed ite Aggregator2 Waiting to read from Transformer (Here the arrows represent dependency direction. this scenario would create a circular dependency where Transformer is waiting on Aggregator1. that “Fork-Join” is a graphical description. not necessarily a Join stage) by Enterprise Edition during job startup. the job would deadlock . Aggregator1 is waiting to write to the Join. and Aggregator2 is waiting on Transformer. It is placed on all inputs to a join structure (again. it doesn’t necessarily have to involve a Join stage.

or translated into any language in any form by any means without the written permission of IBM. All rights reserved. BufferOps will also be placed on the input partitions to any sequential stage that is fed by a parallel stage. as a nonnegative (proper or improper) fraction of APT_BUFFER_MAXIMUM_MEMORY. thus breaking the circular dependency and guaranteeing no deadlock will occur. 2006 137 of 179  Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Values greater than 1 indicate that the buffer July 17. BufferOps will allocate 3MB of memory each (remember that this is per operator. TIP: For very wide rows. When that is full (because the upstream operator is still writing but the downstream operator isn’t ready to accept that data yet) it will begin to flush data to the scratchdisk resources specified in the configuration file (detailed in Chapter 11. stored in a retrieval system. Join cannot be ‘stuck’ waiting to read from either of its inputs. May not exceed 2/3 of APT_BUFFER_MAXIMUM_MEMORY. transmitted. it may be necessary to increase the default buffer size (APT_BUFFER_MAXIMUM_MEMORY) to hold more rows in memory. Default 1 MB. transcribed. The behavior of deadlock-prevention BufferOps can be tuned through these environment variables:  APT_BUFFER_DISK_WRITE_INCREMENT o Controls the “blocksize” written to disk as the memory buffer fills.Information Integration Solutions Center of Excellence BufferOp1 Aggregator 1 Transformer Join Aggregator2 BufferOp 2 (Here the arrows now represent data-flow. APT_BUFFER_FREE_RUN o Maximum capacity of the buffer operator before it starts to offer resistance to incoming flow. No part of this publication may be reproduced. per partition).) Since BufferOp is always ready to read or write. . not dependency. “The Parallel Engine Configuration File” of the DataStage Manager guide). By default. as these same types of circular dependencies can result from partition-wise Fork-Joins.

APT_BUFFERING_POLICY o Specifies the buffering policy for the entire job. transcribed. transmitted. 2006 138 of 179 © 2006 IBM Information Integration Solutions. buffer size. . No part of this publication may be reproduced. but will almost certainly degrade performance and therefore shouldn’t be used in production job runs. BufferOps also have the effect of “smoothing out” production/consumption spikes. and write increment can be set on a per-stage basis from the Input/ Advanced tab of the stage properties. Default is 3 MB. the default buffering policy is AUTOMATIC_BUFFERING. queue bound.Information Integration Solutions Center of Excellence operator will free run (up to a point) even when it has to write data to disk. All rights reserved. use of this setting is only recommend for advanced users! FORCE_BUFFERING can be used to reveal bottlenecks in a job design during development and performance tuning. the buffer mode. stored in a retrieval system. This allows the job to run at the highest rate possible even when a Parallel Framework Red Book: Data Flow Job Design July 17.  APT_BUFFER_MAXIMUM_MEMORY o Maximum memory consumed by each buffer operator for data storage. as shown in the illustration below: Aside from ensuring that no dead-lock occurs. or translated into any language in any form by any means without the written permission of IBM. When it is not defined or defined to be the null string. Valid settings are:  AUTOMATIC_BUFFERING: buffer as necessary to prevent dataflow deadlocks  FORCE_BUFFERING: buffer all virtual Data Sets  NO_BUFFERING: inhibit buffering on all virtual Data Sets  WARNING: Inappropriately specifying NO_BUFFERING can cause dataflow deadlock during job execution. Additionally. buffer free run.

this is called “pushback”. When attempting to address these mismatches in production/consumption. Choosing which stages to tune buffering for and which to leave alone is as much art as science. instead of globally through environment variable settings. and should be considered among the last resorts for performance tuning. As implied above. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. you may be able identify points in the data flow where a downstream stage is waiting on an upstream stage to produce data. when a buffer has consumed its RAM. etc. then setting the buffering policy to “FORCE_BUFFERING” will cause buffering to occur everywhere. if you do not have force buffering set and APT_BUFFER_FREE_RUN set to at least ~1000. Because of this. Stages upstream/downstream from high-latency stages (such as remote databases. If that doesn’t yield enough of a performance boost (remember to test iteratively. All rights reserved. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence downstream stage is ready for data at different times than when its upstream stage is ready to produce that data.) are a good place to start. transmitted. as some other stage far downstream could be responsible for cascading pushback all the way upstream to the place you are seeing the bottleneck. Each place may offer an opportunity for buffer tuning. while changing only 1 thing at a time). By using the performance statistics in conjunction with this buffering. it is best to tune the buffers on a per-stage basis. No part of this publication may be reproduced. NFS mount points for data storage. 2006 139 of 179 © 2006 IBM Information Integration Solutions. you cannot determine that any one stage is waiting on any other stage. . it will ask the upstream stage to “slow down” . stored in a retrieval system.

database login properties…)  environment variables and their default settings (as outlined in Section 2. Development Guidelines Modular development techniques should be used to maximize re-use of DataStage jobs and components.  Create a standard directory structure outside of the DataStage project directory for source and target files.5: Job Parameters. transcribed. especially for DataStage Project categories. as well as Annotation fields. An example directory naming structure is given in Section 2. stage names. See Section 2. using the Manager DSX export capability. DataStage Template jobs should be created with:  standard parameters (for example. user.  A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN. as outlined in Section 3:Development Guidelines:  Job parameterization allows a single job design to process similar logic instead of creating multiple copies of the same job. No part of this publication may be reproduced. . Parallel Framework Red Book: Data Flow Job Design July 17.5.1 Environment Variable Settings)  annotation blocks Job Parameters should always be used for file paths. these standard parameters and settings should be made part of a Designer Job Template. source and target file paths. password. All rights reserved.Information Integration Solutions Center of Excellence Appendix A: Standard Practices Summary This Appendix summarizes Standard Practices recommendations outlined in this document.  Where possible.4: Working with Source Code Control Systems. See Section 2. Standards It is important to establish and follow consistent standards in:  Directory structures for install and application support directories. To ease re-use.  Naming conventions. The scope of a parameter is discussed further in Section 3. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. This can also be used for integration with source code control systems. and links. file names. transmitted.3: Documentation and Annotation. etc) and directories where files are stored. An example DataStage naming structure is given in Section 2. along with cross-references for more detail. and so forth. 2.2: Naming Conventions. or translated into any language in any form by any means without the written permission of IBM. database login settings. stored in a retrieval system. 1. 2006 140 of 179 © 2006 IBM Information Integration Solutions. All DataStage jobs should be documented with Short Description field. intermediate work files. It is the DataStage developer’s responsibility to make personal backups of their work on their local workstation.1:Directory Structures. create re-usable components such as parallel shared containers to encapsulate frequently-used logic.

Objective 2: The partition method must match the business requirements and stage functional requirements. transmitted. BASIC Routines are appropriate only for job control sequences. . Component Usage As discussed in Section 3.  Use the Copy stage as a placeholder for iterative design. Leverage default type conversions using the Copy stage or across the Output mapping tab of other stages. 3.Information Integration Solutions Center of Excellence Parallel Shared Containers should be used to encapsulate frequently-used logic. No part of this publication may be reproduced. stored in a retrieval system. This ensures that the processing workload is evenly balanced. assigning related records to the same partition if required Any stage that processes groups of related records (generally using one or more key columns) must be partitioned using a keyed partition method. 5. the following guidelines should be followed when constructing parallel jobs in DS/EE:  Never use Server Edition components (BASIC Transformer. Standardized Error Handling routines should be followed to capture errors and rejects. the following objectives help to form a methodology for assigning partitioning: Objective 1: Choose a partitioning method that gives close to an equal number of rows in each partition.  Use BuildOp stages only when logic cannot be implemented in the parallel Transformer. Parallel Framework Red Book: Data Flow Job Design July 17.  Use the parallel Transformer stage (not the BASIC Transformer) instead of the Filter or Switch stages. using RCP to maximize re-use. Further details are provided in Section 3. 4. as outlined in Section 4:DataStage Data Types. transcribed. while minimizing overhead.8: Component Usage. 2006 141 of 179 © 2006 IBM Information Integration Solutions.  Always use parallel Data Sets for intermediate storage between jobs. and to facilitate default type conversions. Partitioning Data Given the numerous options for keyless and keyed partitioning. or translated into any language in any form by any means without the written permission of IBM.7:Error and Reject Record Handling. minimizing overall run time.   DataStage Data Types Be aware of the mapping between DataStage (SQL) data types and the internal DS/EE data types. All rights reserved. Server Shared Containers) within a parallel job.

2006 142 of 179 © 2006 IBM Information Integration Solutions. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. Collecting Data Given the options for collecting data into a sequential stream.Information Integration Solutions Center of Excellence This includes. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. Merge. but is not limited to: Aggregator. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. No part of this publication may be reproduced. Join. This is particularly useful if downstream jobs are run with the same degree of parallelism (configuration file) and require the same partition and sort order. stored in a retrieval system. persistent Data Sets can be used to retain the partitioning and sort order. and Sort stages. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). transcribed. Remove Duplicates. Note that in satisfying the requirements of this second objective. use Round Robin partitioning to redistribute data equally across all partitions o Especially useful if the input Data Set is highly skewed or sequential d) Use Same partitioning to optimize end-to-end partitioning and to minimize repartitioning o Being mindful that Same partitioning retains the degree of parallelism of the upstream stage o Within a flow. 6. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. Change Apply. Using the above objectives as a guide. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. Objective 3: Unless partition distribution is highly skewed. minimize repartitioning. Objective 4: Partition method should not be overly complex The simplest method that meets the above objectives will generally be the most efficient and yield the best performance. Change Capture. transmitted. use Auto partitioning (the default) Parallel Framework Red Book: Data Flow Job Design July 17. . the following methodology can be applied: a) Start with Auto partitioning (the default) b) Specify Hash partitioning for stages that require groups of related records o Specify only the key column(s) that are necessary for correct grouping as long as the number of unique values is sufficient o Use Modulus partitioning if the grouping is on a single integer key column o Use Range partitioning if the data is highly skewed and the key column values and distribution do not change significantly over time (Range Map can be reused) c) If grouping is not required. It may also be necessary for Transformers and BuildOps that process groups of related records. Across jobs.

transcribed. All rights reserved. the following methodology should be applied when sorting in a DataStage Enterprise Edition data flow: a) Start with a link sort b) Specify only necessary key column(s) c) Don’t use Stable Sort unless needed d) Use a stand-alone Sort stage instead of a Link sort for options that not available on a Link sort:  Sort Key Mode. The Lookup stage is most appropriate when reference data is small enough to fit into available memory. Merge. sequential ordered result set use a parallel Sort and a Sort Merge collector 8. stored in a retrieval system. transmitted. 7. Create Key Change Column. as long as the Data Set has not been repartitioned or reduced. Join vs. Sorting Using the rules and behavior outlined in Section 6: Sorting. use Sort Merge collector to produce a single. or translated into any language in any form by any means without the written permission of IBM. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. If the Data Sets are larger than available memory resources. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. the Ordered collector may be more efficient c) Use a Round Robin collector to reconstruct rows in input order for round-robin partitioned input Data Sets. Output Statistics  Always specify “DataStage” Sort Utility for standalone Sort stages  Use the “Sort Key Mode=Don’t Sort (Previously Sorted)” to resort a sub-grouping of a previously-sorted input Data Set e) Be aware of automatically-inserted sorts  Set $APT_SORT_INSERTION_CHECK_ONLY to verify but not establish required sort order f) Minimize the use of sorts within a job flow g) To generate a single. . precautions must be taken when using expressions or derivations on nullable columns within the parallel Transformer: o Always convert nullable columns to in-band values before using them in an expression or derivation.1. See Section 9.Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel.  Parallel Framework Red Book: Data Flow Job Design July 17.1: Transformer NULL Handling and Reject Link. No part of this publication may be reproduced.1: Lookup vs.  Stage-Specific Guidelines As discussed in Section 8. Create Cluster Key Change Column. o Always place a reject link on a parallel Transformer to capture / audit possible rejects. 2006 143 of 179 © 2006 IBM Information Integration Solutions. use the Join or Merge stage.

DB2. stored in a retrieval system. Care must be taken to observe the data type mappings documented in Section 10: Database Stage Guidelines when designing a parallel job with DS/EE. use a SQL where clause to limit the number of rows sent to a DataStage job. 2006 144 of 179 © 2006 IBM Information Integration Solutions.1. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. (See Section 9.2Capturing Unmatched Records from a Join). For maximum scalability and parallel performance. 10. it is best to implement business rules natively using DataStage parallel components. 9. transmitted. use the native parallel database stages for maximum performance and scalability. Parallel Framework Red Book: Data Flow Job Design July 17. Use Hash method Aggregators only when the number of distinct key column values is small. the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records. transcribed. Even if the source data is not nullable.   Troubleshooting and Monitoring Always test DS/EE jobs with a parallel configuration file ($APT_CONFIG_FILE) that has two or more nodes in its default pool. Check the Director log for warnings. or Informix databases.  Database Stage Guidelines Where possible. use orchdbutil to properly import design metadata.Information Integration Solutions Center of Excellence    Limit the use of database Sparse Lookups to scenarios where the number of input rows is significantly smaller (for example 1:100 or more) than the number of reference rows. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. All rights reserved. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. . or when exception processing. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow. A Sort method Aggregator should be used when the number of distinct key values is large or unknown. If possible.1: Database stage types: Native Parallel Database Stages DB2/UDB Enterprise Informix Enterprise ODBC Enterprise Oracle Enterprise SQL Server Enterprise Teradata Enterprise      The ODBC Enterprise stage should only be used when a native parallel stage is not available for the given source or target database. as discussed in Section 10. which may indicate an underlying problem or data type conversion issue. When using Oracle.

2 Understanding the Parallel Job Score. 2006 145 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. Enable $APT_DUMP_SCORE by default. All rights reserved. and examine the job score by following the guidelines outlined in Section 12. transcribed. Set the environment variable $OSH_PRINT_SCHEMAS to capture actual runtime schema to the Director log. . Set $DS_PX_DEBUG if the schema record is too large to capture in a Director log entry. transmitted.4. error and warning messages from a running DS/EE job. Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence    The environment variable $DS_PX_DEBUG can be used to capture all generated OSH. stored in a retrieval system. No part of this publication may be reproduced.

Subject Modifier. stored in a retrieval system. transmitted. . No part of this publication may be reproduced. All rights reserved. transcribed. Class Word where the following frequently-used Class Words describe the object type. File. DB) Message (Sequence) Get (Shared Container) Put (Shared Container) Input Output Delete Insert Update Data Store Database Stored Procedure Table View Dimension Fact Source Target Development / Debug Stages Column Generator Head Peek Row Generator Sample Src<job> Load<job> <job>_Seq <job>Psc <job>Ssc <name>Parm Ref Rej Msg Get Put In Out Del Ins Upd DB SP Tab View Dim Fact Src Tgt CGen Head Peek RGen Smpl Tail File Stages Sequential File Complex Flat File File Set Parallel Data Set Lookup File Set External Source External Target Parallel SAS Data Set Processing Stages Aggregator Change Apply Change Capture Copy Filter Funnel Join (Inner) Join (Left Outer) Join (Right Outer) Join (Full Outer) Lookup Merge Modify Pivot Remove Duplicates SAS processing Sort Surrogate Key Generator Switch Transformer Stage Transformer (native parallel) BASIC Transformer (Server) Stage Variable Real Time Stages RTI Input RTI Output XML Input July 17. 2006 Tail SF CFF FS DS LFS XSrc XTgt SASd Agg ChAp ChCp Cp Filt Funl InJn LOJn ROJn FOJn Lkp Mrg Mod Pivt RmDp SASp Srt SKey Swch Tfm BTfm SV RTIi RTIo XMLi 146 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. or the function the object performs: Project Repository and Components <proj>Dev Development <proj>Test Test <proj>Prod Production BdOp<name> BuildOp XFn<name> Parallel External Function Wrap<name> Wrapper Job Names and Properties Extract Job Load Sequence Parallel Shared Container Server Shared Container Parameter Links (prefix with “lnk_”) Reference (Lookup) Reject (Lookup.Information Integration Solutions Center of Excellence Appendix B: DataStage Naming Reference Every name should be based on a three-part concept: Subject. or translated into any language in any form by any means without the written permission of IBM.

No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. transcribed. 2006 147 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. . All rights reserved.Information Integration Solutions Center of Excellence XML Output XML Transformer Restructure Stages Column Export Column Import XMLo XMLt CExp CImp Parallel Framework Red Book: Data Flow Job Design July 17. transmitted.

the job score is output to the DataStage Director log.Information Integration Solutions Center of Excellence Appendix C: Understanding the Parallel Job Score Jobs developed in DataStage Enterprise Edition are independent of the actual hardware and degree of parallelism used to run the job. and the overhead to capture the score is negligible. The first score is from the license operator. It is recommended that this setting be enabled by default at the project level. as the job score offers invaluable data for debugging and performance tuning. the DS/EE job score: • • • • • Identifies degree of parallelism and node assignment(s) for each operator Details mappings between functional (stage/operator) and actual operating system processes Includes operators automatically inserted at runtime: o Buffer operators to prevent deadlocks and optimize data flow rates between stages o Sorts and Partitioners that have been automatically inserted to ensure correct results Outlines connection topology (Data Sets) between adjacent operators and/or persistent Data Sets Defines number of actual operating system processes Where possible. degree of parallelism and node (server) assignments. At runtime. multiple operators are combined within a single operating system process to improve performance and optimize resource requirements. transcribed. 1. The parallel Configuration File provides a mapping at runtime between the compiled job and the actual runtime infrastructure and resources by defining logical processing nodes. As shown in the illustration below. not the actual job. or translated into any language in any form by any means without the written permission of IBM. . transmitted. Viewing the Job Score When the environment variable APT_DUMP_SCORE is set. Parallel Framework Red Book: Data Flow Job Design July 17. Similar to the way a parallel database optimizer builds a query plan. job score entries start with the phrase “main_program: This step has n datasets…” Two separate scores are written to the log for each job run. 2006 148 of 179 © 2006 IBM Information Integration Solutions. Enterprise Edition uses the given job design and configuration file to compose a job score which details the processes created. No part of this publication may be reproduced. All rights reserved. and can be ignored. The second score entry is the actual job score. stored in a retrieval system. and interconnects (Data Sets) between them.

No part of this publication may be reproduced. All rights reserved. In this example. In this example. there are two virtual Data Sets. links between stages). 2006 149 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. for a total of 9 operating system process. two running in parallel across 4 nodes. July 17.Information Integration Solutions Center of Excellence Actual job score 2. Terminology in this section can be used to identify the type of partitioning or collecting that was used between operators. transcribed. and the degree of parallelism per operator o Node assignment for each operator. there are 3 operators. • Operators: starts with the words “It has n operators:” The second section details actual operators created to execute the job flow. as shown in the example on the right: • Data Sets: starts with the words “main_program: This step has n datasets:” The first section details all Data Sets. including persistent (on disk) and virtual (in memory. “node3”. transmitted. This includes: o Sequential or Parallel operation. (in this example: “node1”. Parallel Job Score Components The Enterprise Edition parallel job score is divided into two sections. “node2”. “node4”). or translated into any language in any form by any means without the written permission of IBM. one running sequentially. stored in a retrieval system. The actual node names correspond to node names in the parallel configuration file. .

. the first Data Set is identified as “ds0”. Within the Data Set definition. or translated into any language in any form by any means without the written permission of IBM. 3. Job Score: Data Sets The parallel pipeline architecture of DataStage Enterprise Edition passes data from upstream producers to downstream consumers through in-memory virtual data sets. Producer Partitioner Collector Consumer The notation between producer and consumer is used to report the type of partitioning or collecting (if any) that is applied. collector type with the second. For example. The symbol between the partition name and collector name indicates: -> <> => #> >> > Sequential producer to Sequential consumer Sequential producer to Parallel consumer Parallel producer to Parallel consumer (SAME partitioning) Parallel producer to Parallel consumer (repartitioned. while operators are identified by their operator number and name. Producers and consumers may be either persistent (on disk) Data Sets or parallel operators. operator zero (op0) is running sequentially. The memory used by deadlock-prevention BufferOps can be calculated based on the number of inserted BufferOps. the upstream producer is identified first. followed by the downstream consumer. for persistent Data Sets) July 17.Information Integration Solutions Center of Excellence Note that the number of virtual Data Sets and the degree of parallelism determine the amount of memory used by the inter-operator transport buffers. corresponding to the lower section of the job score. 2006 150 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. transmitted. All rights reserved. not SAME) Parallel producer to Sequential consumer No producer or no consumer (typically. followed by a notation to indicate the type of partitioning or collecting (if any). The partition type is associated with the first term. Persistent Data Sets are identified by their Data Set name. stored in a retrieval system. and the next “ds1”. Operator 1 (op1) is running in parallel with 4 degrees of parallelism [4p]. with 1 degree of parallelism [1p]. Data Sets are identified in the first section of the parallel job score. transcribed. as illustrated in the example on the right: The degree of parallelism is identified in brackets after the operator name. No part of this publication may be reproduced. in the example on the right. with each Data Set identified by its number (starting at zero). In the above example.

p2] ecc3674[op3. a composite operator includes more than one function. or translated into any language in any form by any means without the written permission of IBM.p0] from zero (example “op0”) )} op1[4p] {(parallel inserted tsort operator • degree of parallelism within brackets {key={value=LastName}. each individual component of a composite operator is represented as an individual operator in the job score.p0] ecc3672[op4.Information Integration Solutions Center of Excellence Finally.p3] )} op4[4p] {(parallel APT_CombinedOperatorController: (APT_LUTProcessImpl in Lookup_3) (APT_TransformOperatorImplV0S7_cpLookupTest1_Tran sformer_7 in Transformer_7) (PeekNull) ) on nodes ( ecc3671[op4. stored in a retrieval system.p2] node4[op2.p0] )} . Job Score: Operators The lower portion of the parallel job score details op0[1p] {(sequential the mapping between stages and actual processes APT_CombinedOperatorController: (Row_Generator_0) generated at runtime. (example “[4p]”) key={value=FirstName}}(0)) on nodes ( • “sequential” or “parallel” execution mode node1[op2. includes (as illustrated in the job score fragment): key={value=FirstName}}) ) on nodes ( • operator name (opn) numbered sequentially node1[op0. 4.p0] data has been loaded ecc3672[op3. if the Preserve Partitioning flag has been set for a particular Data Set. For each operator. transmitted.p0] o may also include combined node2[op2. For example. a composite operator appears to be a single stage on the design canvas. Lookup) Some stages are composite operators – to the DataStage developer.p1] At runtime.p3] )} specified stage name in the Designer op2[4p] {(parallel buffer(0)) canvas on nodes ( node1[op2.p0] node2[op2. No part of this publication may be reproduced.APT_LUTProcessImpl: op3[4p] {(parallel buffer(0)) Performs actual lookup processing once reference on nodes ( ecc3671[op3.p3] (APT_CombinedOperatorController) which include logic from multiple stages in a single operator o may also include framework-inserted operators such as Buffers. All rights reserved. . transcribed. this (inserted tsort operator {key={value=LastName}. Since Parallel Framework Red Book: Data Flow Job Design July 17. as shown in the following score fragment shown on the right: ecc3673[op3. 2006 151 of 179 © 2006 IBM Information Integration Solutions. Lookup is a composite operator.APT_LUTCreateImpl: op2[1p] {(parallel APT_LUTCreateImpl in Lookup_3) on nodes ( Reads the reference data into memory ecc3671[op2.p1] operators node3[op2.p1] ecc3673[op4. you can evaluate the memory used by a lookup.p1] • Components of the operator node3[op2. the notation “[pp]” will appear in this section of the job score.p2] Using this information together with the output from the $APT_PM_SHOW_PIDS environment variable.p2] o typically correspond to the usernode4[op2. Sorts o may include “composite” operators (for example. It is composted of the following internal operators: . But internally.

ds)} .ds) .ds} It has 3 operators: op0[1p] {(sequential Row_Generator_0) on nodes ( node1[op0. In a similar way.Delete Data Files ->eCollectAny op2[1p] (sequential delete descriptor file in delete temp. All rights reserved.p0] )} op2[1p] {(sequential delete descriptor file in delete temp.ds) on nodes ( node1[op1. transmitted. stored in a retrieval system.p0] )} op1[1p] {(parallel delete data files in delete temp. 2006 152 of 179 © 2006 IBM Information Integration Solutions. transcribed. you can also determine the delay associated with loading the lookup structure. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced.Delete Descriptor File ds1: {op0[1p] (sequential Row_Generator_0) -> temp.p0] )} Parallel Framework Red Book: Data Flow Job Design July 17.ds) on nodes ( node1[op2.Information Integration Solutions Center of Excellence the entire structure needs to be loaded before actual lookup processing can begin. a persistent Data Set defined to “Overwrite” an existing Data Set of the same name main_program: This step has 2 datasets: will have multiple entries in the job score to: ds0: {op1[1p] (parallel delete data files in delete temp. .

if you have a transfer that transfers all fields from input to output.Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. the method APT_Record::estimateFinalOutputSize() can give you an estimate for a given record schema.add (# nullable fields)/8 for null indicators . transmitted. All rights reserved. rounded up 8 bytes n + 4 bytes for non-NLS data 2n + 4 bytes for NLS data (internally stored as UTF-16) n bytes for non-NLS data 2n bytes for NLS data 4 bytes 8 bytes with microsecond resolution 4 bytes 8 bytes 12 bytes with microsecond resolution For the overall record width: .one byte per column for field alignment (worst case is 3. stored in a retrieval system. As can APT_Transfer::getTransferBufferSize(). this Appendix provides a more accurate and detailed way to estimate the size of a parallel Data Set based on the internal storage requirements for each data type: Data Type Integers Small Integer Tiny Integer Big Integer Decimal Float VarChar(n) Char(n) Time Date Timestamp Size 4 bytes 2 bytes 1 byte 8 bytes (precision+1)/2. transcribed.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 153 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. . No part of this publication may be reproduced.

Defines size of I/O buffer for Sequential File reads (imports) and writes (exports) respectively. Setting this environment variable directs DataStage to reject Sequential File records with strings longer than their declared maximum column length. By default. NOTE: The environment variable settings in this Appendix are only examples.Information Integration Solutions Center of Excellence Appendix E: Environment Variable Reference This Appendix summarizes the environment variables mentioned throughout this document. Setting this value to a low number (such as 1) is useful for realtime applications. July 17. Environment Variable $APT_EXPORT_FLUSH_COUNT Sequential File Stage Environment Variables Setting [nrows] Description Specifies how frequently (in rows) that the Sequential File stage (export operator) flushes its internal buffer to disk. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. . setting this variable to a value equal to the read / write size in bytes can improve performance of Sequential File import/export operations. 2006 154 of 179 $APT_IMPORT_REJECT_STRING_FIELD_OVERRUN S 1 (DataStage v7.2: Default and Explicit Type Conversions 2. with a minimum of 8. transmitted. No part of this publication may be reproduced. Increasing these values on heavily-loaded file servers may improve performance.1. These variables can be used on an as-needed basis to tune the performance of a particular job flow. All rights reserved. a variable-length string field to a fixed length (or a fixed-length to a longer fixed-length). or to change the default behavior of specific DataStage Enterprise Edition stages. transcribed. to assist in debugging. or translated into any language in any form by any means without the written permission of IBM. In some disk array configurations. or pads. stored in a retrieval system. See section 4. 1. Set values that are optimal to your environment. imported string fields that exceed their maximum declared length are truncated. but there is a small performance penalty from increased I/O. Default is 128 (128K).01 and later) [Kbytes] $APT_IMPORT_BUFFER_SIZE $APT_EXPORT_BUFFER_SIZE $APT_CONSISTENT_BUFFERIO_SIZE [bytes] Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Environment Variable $APT_STRING_PADCHAR Job Design Environment Variables Setting [char] Description Overrides the default pad character of 0x0 (ASCII null) used when EE extends.

When more than 500 bytes read-ahead is desired. $DB2DBDFT is used to find the database name. Specifies the name of the DB2 database for DB2/UDB Enterprise stages if the “Use Database Environment Variable” option is True. transcribed. Specifies the number of records to insert between commits. The default value is 2000 per partition.000 bytes. . Allows DataStage plug-in stages to handle DB2 databases which use the special characters # and $ in column names. $DS_ENABLE_RESERVED_CHAR_CONVE RT [rows] 1 Parallel Framework Red Book: Data Flow Job Design July 17. but this can be set as low as 2 bytes. This setting should be set to a lower value when reading from streaming inputs (for example. stored in a retrieval system. 2006 155 of 179 © 2006 IBM Information Integration Solutions. this will significantly improve file pattern reads. $APT_MAX_DELIMITED_READ_SIZE [bytes] $APT_IMPORT_PATTERN_USES_FILESET [set] 3.Information Integration Solutions Center of Excellence $APT_DELIMITED_READ_SIZE [bytes] Specifies the number of bytes the Sequential File (import) stage reads-ahead to get the next delimiter. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. transmitted.000 bytes. By default. socket or FIFO) to avoid blocking. No part of this publication may be reproduced. The default is 500 bytes. This variable controls the upper bound which is by default 100.db2profile. This variable is usually set in a user’s environment from . Used as a backup for specifying the DB2 installation directory (if $INSTHOME is undefined). If it is not found the importer looks ahead 4*500=2000 (1500 more) bytes. use this variable instead of APT_DELIMITED_READ_SIZE. For disk configurations with multiple controllers and disk. If $APT_DBNAME is not defined. Sequential File (import) will read ahead 500 bytes to get the next delimiter. $APT_DB2INSTANCE_HOME $APT_DBNAME $APT_RDBMS_COMMIT_ROWS Can also be specified with the “Row Commit Interval” stage input property. and so on (4X) up to 100. Environment Variable $INSTHOME DB2 Environment Variables Setting [path] [path] [database] Description Specifies the DB2 install directory. When this environment variable is set (present in the environment) file pattern reads are done in parallel by dynamically building a File Set header based on the list of files that match the given expression.

Specifies Oracle SQL*Loader options used in a target Oracle stage with Load method. stored in a retrieval system. Specifies the path to the Informix sqlhosts file. No part of this publication may be reproduced. a target Oracle stage with Load method will limit the number of players to the number of datafiles in the table’s tablespace. . When set. By default. Environment Variable $ORACLE_HOME Oracle Environment Variables Setting [path] [sid] [num] [seconds] Description Specifies installation directory for current Oracle instance. By default. These two environment variables work together to specify how often target rows are committed for target Oracle stages with Upsert method. 5. Informix Environment Variables Environment Variable $INFORMIXDIR $INFORMIXSQLHOSTS $INFORMIXSERVER $APT_COMMIT_INTERVAL Setting [path] [filepath] [name] [rows] Description Specifies the Informix install directory. Normally set in a user’s environment by Oracle scripts. Specifies the name of the Informix server matching an entry in the sqlhosts file. The filepath specified by this environment variable specifies the file with the SQL*Loader commands. 2006 156 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. commits are made every 2 seconds or 5000 rows per partition. or translated into any language in any form by any means without the written permission of IBM. The default is 10000 per partiton. July 17.Information Integration Solutions Center of Excellence 4. this is set to OPTIONS(DIRECT=TRUE. transcribed. Specifies the Oracle service name. Specifies the commit interval in rows for Informix HPL Loads. Commits are made whenever the time interval period has passed or the row interval is reached. Setting this variable makes it possible to load fields with trailing or leading blank characters. All rights reserved. When set. transmitted. corresponding to a TNSNAMES entry. PARALLEL=TRUE) $ORACLE_SID $APT_ORAUPSERT_COMMIT_ROW_INTERVAL $APT_ORAUPSERT_COMMIT_TIME_INTERVAL $APT_ORACLE_LOAD_OPTIONS [SQL* Loader options] [char] $APT_ORACLE_LOAD_DELIMITED (DataStage 7.01 and later) $APT_ORA_IGNORE_CONFIG_FILE_PARALLELIS M 1 $APT_ORA_WRITE_FILES [filepath] Specifies a field delimiter for target Oracle stages using the Load method. the output of a Target Oracle stage with Load method is written to files instead of invoking the Oracle SQL*Loader. Useful in debugging Oracle SQL*Loader issues. whichever comes first.

Information Integration Solutions Center of Excellence
$DS_ENABLE_RESERVED_CHAR_CONVERT

1 Allows DataStage plug-in stages to handle Oracle databases which use the special characters # and $ in column names.

6.
Environment Variable
$APT_TERA_SYNC_DATABASE $APT_TERA_SYNC_USER $APT_TER_SYNC_PASSWORD $APT_TERA_64K_BUFFERS

Teradata Environment Variables
Setting
[name] [user] [password] 1 1

Description
Starting with v7, specifies the database used for the terasync table. By default, EE uses the Starting with v7, specifies the user that creates and writes to the terasync table. Specifies the password for the user identified by $APT_TERA_SYNC_USER. Enables 64K buffer transfers (32K is the default). May improve performance depending on network configuration. This environment variable is not recommended for general use. When set, this environment variable may assist in job debugging by preventing the removal of error tables and partially written target table. Disables permission checking on Teradata system tables that must be readable during the TeraData Enterprise load process. This can be used to improve the startup time of the load.

$APT_TERA_NO_ERR_CLEANUP

$APT_TERA_NO_PERM_CHECKS

1

7.
Environment Variable
$APT_MONITOR_TIME

Job Monitoring Environment Variables
Setting [seconds] Description In v7 and later, specifies the time interval (in seconds) for generating job monitor information at runtime. To enable size-based job monitoring, unset this environment variable, and set $APT_MONITOR_SIZE below. Determines the minimum number of records the job monitor reports. The default of 5000 records is usually too small. To minimize the number of messages during large job runs, set this to a higher value (for example, 1000000). Disables job monitoring completely. In rare instances, this may improve performance. In general, this should only be set on a per-job basis when attempting to resolve performance bottlenecks. Prints record counts in the job log as each operator completes processing. The count is per operator per partition.

$APT_MONITOR_SIZE

[rows]

$APT_NO_JOBMON

1

$APT_RECORD_COUNTS

1

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

157 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

8. PerformanceTuning Environment Variables
Environment Variable
$APT_BUFFER_MAXIMUM_MEMORY

Setting 41903040 (example)

$APT_BUFFER_FREE_RUN

1000 (example)

Description Specifies the maximum amount of virtual memory, in bytes, used per buffer per partition. If not set, the default is 3MB (3145728). Setting this value higher will use more memory, depending on the job flow, but may improve performance. Specifies how much of the available in-memory buffer to consume before the buffer offers resistance to any new data being written to it. If not set, the default is 0.5 (50% of $APT_BUFFER_MAXIMUM_MEMORY). If this value is greater than 1, the buffer operator will read $APT_BUFFER_FREE_RUN * $APT_BUFFER_MAXIMIMUM_MEMORY before offering resistance to new data. When this setting is greater than 1, buffer operators will spool data to disk (by default scratch disk) after the $APT_BUFFER_MAXIMUM_MEMORY threshold. The maximum disk required will be
$APT_BUFFER_FREE_RUN * # of buffers * $APT_BUFFER_MAXIMUM_MEMORY

$APT_PERFORMANCE_DATA

directory [path]

$TMPDIR

Enables capture of detailed, per-process performance data in an XML file in the specified directory. Unset this environment variable to disable. Defaults to /tmp. Used for miscellaneous internal temporary data including FIFO queues and Transformer temporary storage. As a minor optimization, may be best set to a filesystem outside of the DataStage install directory.

9.
Environment Variable
$OSH_PRINT_SCHEMAS

Job Flow Debugging Environment Variables
Setting
1

Description
Outputs the actual schema definitions used by the DataStage EE framework at runtime in the DataStage log. This can be useful when determining if the actual runtime schema matches the expected job design table definitions. Disables operator combination for all stages in a job, forcing each EE operator into a separate process. While not normally needed in a job flow, this setting may help when debugging a job flow or investigating performance by isolating individual operators to separate processes. Note that disabling operator combination will generate more UNIX processes, and hence require more system resources (and memory). Disabling operator July 17, 2006 158 of 179

$APT_DISABLE_COMBINATION

1

The Advanced Stage Properties editor in DataStage Designer v7.1 and later allows combination to be enabled and disabled for on a per-stage basis.

Parallel Framework Red Book: Data Flow Job Design

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

$APT_PM_PLAYER_TIMING

1 1 FORCE

$APT_PM_PLAYER_MEMORY $APT_BUFFERING_POLICY

combination also disables internal optimizations for job efficiency and run-times. Prints detailed information in the job log for each operator, including CPU utilization and elapsed processing time. Prints detailed information in the job log for each operator when allocating additional heap memory. Forces an internal buffer operator to be placed between every operator. Normally, the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. Using the job monitor performance statistics, this can identify which part of a job flow is impacting overall performance. Set this environment variable to capture copies of the job score, generated osh, and internal Enterprise Edition log messages in a directory corresponding to the job name. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. This environment variable should not normally need to be set. When trying to start very large jobs on heavilyloaded servers, lowering this number will limit the number of processes that are simultaneously created when a job is started. For heavily loaded MPP or clustered environments, this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. The default is 30 seconds.

Setting
$APT_BUFFERING_POLICY=FORCE is not

recommended for production job runs.
$DS_PX_DEBUG

1

$APT_PM_STARTUP_CONCURRENCY

5

$APT_PM_NODE_TIMEOUT

[seconds]

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

159 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

e.k. record adjacency. at the same time. It will be followed by a detailed example that discusses these ideas in much greater depth. not globally. This creates partition-wise concurrency (a. transmitted. key clustering. but there is little you Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. Merge. it does not care about the key value of the next cluster with respect to the current key value—in part because this stage takes only one input. RemoveDuplicates.k. b) Hash gathers Hash gathers into the same partition. sorting is needed to satisfy an input requirement of a downstream stage. but other rows may separate them within that partition. Join. 2006 160 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. key-clustering is sufficient in many cases (a-only.) on any super-set of the keys. a) Sort within partitions. for example. in order This Appendix contains descriptions of what happens “behind the scenes”. etc. An illustrative piece of information. All rights reserved. rows from all partitions that share the same value in key columns. for example. stored in a retrieval system.Information Integration Solutions Center of Excellence Appendix F: Sorting and Hashing Advanced Example The standard recipe for using the ‘Inter-Record Relationship Suite’ (Sort. but is frequently inefficient as records are ‘over-hashed’ and ‘over-partitioned’. this may be review for you. for example. Sort actually does two things: (i) Groups rows that share the same values in key columns (forces related rows to be contiguous. Sorting is rarely required by the business logic. in the same order.a. Remove Duplicates requires only (i): when it completes processing all the rows in a key cluster. and related stages) is: Hash and Sort/Join/Merge on exactly the same keys. The reason for this requirement lies in the “light-weight” nature of these stages. transcribed. Join. If you have a lot of experience with hashing and sorting. Join. This approach is guaranteed to work. These operations take place in parallel. below). There is also an “advanced” rule: a) b) Hash on any sub-set of the keys Sort (join/etc. Merge. only needs to see two records at a time —one from each input stream—to do its job.a. a. all require sorted inputs. c) Sort clusters and orders Sorting is often overkill. across all partitions.) (ii) Orders the clusters resulting from (i). In most cases. . related rows are in the same partition. RemoveDuplicates. partition-wise co-location). i. The second portion of this Appendix assumes you have read and thoroughly understand these concepts. will notice that two rows have identical values in the user-defined key column only if the two rows are contiguous.

This due. ‘streaming-style’). To restore row-adjacency. There is a component that will allow you to partition sorted data and achieve a sorted result: parallelsortmerge (PSM). grouped/clustered for remove duplicates). in part. Nonetheless. a sort operation is needed even on previously sorted columns following any partitioner. Partitioners. On multi-partitioned (i. any existing sort order is usually destroyed—see example below. See usage notes in footnotesiii. Note that neither partition has a sorted result despite P1 having a sorted input (read row order from the bottom up. for example: Repartitioning: P0 2 1 3 P1 103 102 101 Note that ‘1’ and ‘101’ have switched partitions. but it can be invoked via the generic stage. Enterprise Edition itself normally manages this use of this component. so row order between the two inputs is obviously critical. non-sequential) inputs. one from each input. Join and Merge.. they DO care about the key value of the next cluster. work in parallel. In other words. partitioners (except for SAME). to the fact these stages take multiple input links—these stages only see 2 records at a timei. (and most other stages) do not gratuitously alter existing intra-partition row order. partitioners.e. clustering. and your data will retain its previous sort order. When this stage completes processing all the rows in a key cluster. reshuffle rows among partitions. stored in a retrieval system. If the values on both inputs aren't ordered (vs. d) Partitioners respect row order but split clusters. transcribed. All rights reserved. As you will see in this Appendix. No part of this publication may be reproduced. Join/Merge can't effectively choose which input stream to advance for finding subsequent key matches. 1. like stages. this is one instance where this might be handy). . 2006 161 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. however. allow a row in a partition to jump ahead of another row in the same partitionii. there are more advanced methods to sort and partition that can leverage this capability and mitigate the cost of sorting vs. Partitioner P0 2 101 P1 1 103 July 17. as a rule. require both (i) and (ii). transmitted. Enterprise Edition will not. Inside a partitioner In Enterprise Edition. on the other hand. Whenever you re-partition your sorted data. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence can do to take advantage of it as there are no stages which guarantee key-clustering but do not perform a sort (some databases might be able to do key-clustering more cheaply than a sort. follow the partitioner with a PSM.

transcribed. Illustrated Above Before Partitioner Partition 0 3 Eve Smith Pine 3 Eve 1 Orlando Jones Elm 2 Adam 2 Adam Smith Pine 10 Rose 1 Partition 1 10 Rose Jones Pine 1 Orlando 1 10 Boris Smith Walnut 10 Boris 2 2 10 John Zorn Walnut 10 John 3 3 After Hash partitioning on Street/Tree: Illustrated Above After Partitioner Partition 0 Smith 3 Eve Smith Pine 3 Eve Smith 10 Rose Jones Pine 1 Orlando 1 Jones 2 Adam Smith Pine 10 Rose 1 Partition 1 Jones 10 Boris Smith Walnut 10 Boris 2 2 Smith 10 John Zorn Walnut 10 John 3 Zorn3 1 Orlando Jones Elm 2 Adam Smith Jones Jones Smith Zorn Smith There is more than one way to correctly hash-partition any Data Set6. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. Here is another possible outcome: Also: Consider the result of running the same job with the same data. Minimizing Record Movement for Maximizing Performance Now we have covered the basic rules and mechanics for hash-partitioning and sorting. stored in a retrieval system. 6 There is an exception to this rule: If your hash key has only one value.Information Integration Solutions Center of Excellence 3 102 Example: 6 rows in 2 partitions. All rights reserved. . but a different number of partitions. let’s look at how we can capitalize on these behaviors for performance benefits. 2006 162 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Partition 0 Partition 1 1 Orlando 10 Rose 1 10 Boris 2 2 Adam 10 3 3 John Eve Jones Jones Elm Pine 2 1 Adam Smith Orlando Jones Boris Rose John Eve Smith Jones Zorn Smith Smith Walnut 10 2 Smith Pine 10 1 Zorn Walnut 10 3 Smith Pine 3 2. transmitted. July 17.

No part of this publication may be reproduced. thereby increasing the value of this exercise. All rights reserved. as well as the average transaction amount per-item. per store for all stores in the nation. stored in a retrieval system. Here is our source data: Data Set 1: 32 Rows Store Location 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Item ID 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 Transaction Date 2004-01-01 2004-01-02 2004-01-01 2004-01-03 2004-01-04 2004-01-02 2004-01-04 2004-01-03 2004-01-04 2004-01-03 2004-01-01 2004-01-04 2004-01-02 2004-01-02 2004-01-01 2004-01-03 2004-01-03 2004-01-01 2004-01-02 2004-01-04 2004-01-02 2004-01-03 2004-01-01 2004-01-04 2004-01-04 2004-01-01 2004-01-03 2004-01-02 2004-01-04 2004-01-02 2004-01-01 2004-01-03 Transaction Amt 1 2 3 5 5 54 7 8 2 3 45 65 7 85 9 98 23 3 32 45 54 56 7 8 23 45 534 6 65 7 78 87 The screen capture below shows how to implement the business logic in an efficient manner. transcribed. taking advantage of Enterprise Edition’s ability to analyze a jobflow and insert sorts and partitioners in Parallel Framework Red Book: Data Flow Job Design July 17. and append these values to the original data. They wish to determine the weighted average transaction amount peritem nation-wide. . transmitted. There are many common extensions on gathering these kinds of sales metrics that take the following ideas and increase the scale of the problem at hand. or translated into any language in any form by any means without the written permission of IBM. This would make it possible to determine how well each store is doing in relation to the national averages and track these performance trends over time. 2006 163 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets nation-wide.

Information Integration Solutions Center of Excellence appropriate places notice. you will disable this facility. with ANY value. No part of this publication may be reproduced. . If you allow this environment variable to exist with any value. and calculate the average of the ‘Transaction Amt’ column and place the results in a column named ‘Store Average Item Transaction Amt’. To do this. APT_NO_PART_INSERTION is NOT defined in your environment. transmitted. 7 To enable automatic partition insertion. with ANY value. automatic sort insertion. APT_NO_SORT_INSERTION is NOT defined in your environment. Here. Enterprise Edition will hash-partition and sort on ‘Store ID’. The Aggregator stage NationalAverageItemTransactionAmt will aggregate the data on ‘Item ID’. so you want to leave them enabled (the default). both on the output link of JoinSourceToAggregator_1): NOTE: In this job. to get the original data with the averages appended. not store). 2006 164 of 179 © 2006 IBM Information Integration Solutions. ensure that the environment variable. ensure that the environment variable. This is done with JoinSourceToAggregator_1 and JoinSourceToAggregator_2. This is the per-store transaction average per item. or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. we will need to join each aggregator’s output back to the original data in order to get the original row count. All rights reserved. If you allow this environment variable to exist with any value. This is the nation-wide transaction average per item (weighted by transaction. Since the aggregator reduces row count (to the group count). you will disable this facility. and ‘Transaction Date’. and ‘Transaction Date’. To enable automatic sort insertion. stored in a retrieval system. The Aggregator StoreAverageItemTransactionAmt will aggregate the data on ‘Store ID’. ‘Item ID’. transcribed. there is only one sort and one repartition in the diagram. and automatic partition insertion must be turned on7. Here you want to let DS/EE choose where to insert sorts and partitioners for you. and ‘Transaction Date’ calculating the average of the ‘TransactionAmt’ column and place the results in a column named ‘National Average Item Transaction Amt’. Enterprise Edition will hash-partition and sort on ‘Item ID’. ‘Item ID’. and ‘Transaction Date’.

The hash method only requires the input data to be hashed. transcribed.5 50. stored in a retrieval system. It does this by keeping running totals in memory for the aggregation for each output group. No part of this publication may be reproduced.5 35.5 23 23 23 23 26.5 39. it does not guarantee output order. 2006 165 of 179 1 1 2 2 1 1 1 1 1 1 1 1 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 1 3 3 7 7 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.25 26. The sort method requires the input data to be hashed and sorted. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 180.5 310. however.25 26. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.5 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 56 23 85 7 6 7 3 98 87 534 PeekFinalOutput.5 16.5 6.5 46 46 6.5 35.5 180.25 180.5 3.5 39.Information Integration Solutions Center of Excellence NOTE: For this scenario. All rights reserved.25 26.25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17.5 310. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation. or translated into any language in any form by any means without the written permission of IBM. and ‘Average Item Transaction Amt By Store’) The output Data Set should look something like this (A 3-node configuration file was used in this implementation): Data Set 2: 32 Rows PeekFinalOutput.5 180.25 16.5 3.5 50.5 3.5 35. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’. in return. you will need to set the Aggregator’s “Method” to Sort.5 6. Therefore. . it guarantees the output order to be sorted since the result of aggregation can be released for downstream processing as soon as the key change is detected. not Hash.5 Store Average Item Transaction Amt 28 28 43 43 6. transmitted.

and 1.5 27 27 61. appended to the end of this document for masochistsiv) might suggest: 4. A deeper examination (of the score dump. .: record adjacency assumes partition concurrency8. stored in a retrieval system.5 44 44 Since both the Aggregator and Join expect the data to arrive hashed and sorted on the grouping key(s) —both operations that consume large amounts of CPU—a couple of questions arise with respect to efficiency: What is the minimum number of hash-partitioners needed to implement this solution correctly? What is the minimum number of sorts needed to implement this solution? What is the minimum number of times that sort will need to buffer the entire Data Set to disk to implement this solution? Though running the job sequentially eliminates questions related to partitioners.75 38.5 33. All rights reserved. i.25 44.25 38. This is certainly an improvement on the previous answer. 6. transmitted.75 38. as only partition concurrency is affected by sequential execution.25 44. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17.25 44.5 33.Information Integration Solutions Center of Excellence 2 2 1 1 2 2 1 1 2 2 1 1 2 2 2 2 2 2 2 2 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 8 45 45 9 78 45 2 65 65 23 16. Here’s a screen shot of this.e.25 44. or translated into any language in any form by any means without the written permission of IBM.5 61. 2006 166 of 179 © 2006 IBM Information Integration Solutions. more efficientv.25 16. and 6.75 26. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition. and 3. 3. An examination of the job above would suggest: 6. even sequential job execution does not alter the answer for the sort-related questions. 3. A much better answer is: 1.75 38. No part of this publication may be reproduced.

but we know that the data is already sorted on TransactionDate. and automatic partition insertion must be turned offvii. we still need to fix the sort order. All rights reserved. as it isn’t available on the link sort. the data is not properly prepared for StoreAverageItemTransactionAmt. transcribed. as it expects all of the records for a particular StoreLocation/TransactionDate/ItemId combination to arrive on the same partition. 2006 167 of 179 © 2006 IBM Information Integration Solutions. The settings in the sort should look like this: Parallel Framework Red Book: Data Flow Job Design July 17. and ItemId. transmitted. keys (viii This footnote contains key concepts that this document addresses. stored in a retrieval system. No part of this publication may be reproduced. just as in the previous example. TransactionDate. but it is a lengthy parenthetical statement that would interrupt the flow of the scenario discussion). In our initial copy stage (DistributeCopiesOfSourceData). This is because the data is already partitioned in a compatible manner for the aggregator. This combination of hash and sort adequately prepares the data for NationalAverageItemTransactionAmt. automatic sort insertion. The ‘advanced’ rule for hash partitioning is: you may partition on any sub-set of the aggregation/join/sort/etc. we hash and sort on ItemID and TransactionDate only. This is a problem for StoreAverageItemTransactionAmt. Hashing on these fields will gather all unique combinations of ItemID and TransactionDate into the same partition.Information Integration Solutions Center of Excellence NOTE: In this job. What is wrong with the data? The sort order does not include the StoreLocation. However. Sort offers an efficiency-mode for pre-sorted data. but you must use the sort stage to access it. One would expect that we would need to sort on StoreLocation. However. in order. and ItemId. . You may be wondering why the partitioning wasn’t mentioned as part of the problem. or translated into any language in any form by any means without the written permission of IBM.

but the group keys will force the proper order). especially in MPP environments where repartitioning implies network I/O). This means that the output of the DistributeCopiesOfSourceData and NationalAverageItemTransactionAmt are already hashed and sorted on the keys needed to perform JoinSourceToAggregator_1. however dreadful). Getting back to the aggregators. we have instructed the sort stage that the data is already sorted on ItemID and TransactionDate (as always with sorting records.Information Integration Solutions Center of Excellence As you can see. the rows will come out in the same order they went in (different rows. key order is very important). but the entire Data Set was 100 million records. However since we sub-sorted the data before this aggregator (unlike NationalAverageItemTransactionAmt). If the group size was only several hundred records. This lets sort know that it only needs to gather all records with a unique combination of ItemID and TransactionDate in order to sort a batch of records. ALL extant values of the remaining columns are already in the same partition. This accomplishes the first goal. we will have to prep the output from the first join to account for Parallel Framework Red Book: Data Flow Job Design July 17. 2006 168 of 179 © 2006 IBM Information Integration Solutions. to append a column representing the national (weighted) average item transaction amount. No part of this publication may be reproduced. The previous two paragraphs contain two key concepts in Enterprise Edition (pun fully intended. The output of StoreAverageItemTransactionAmt contains the other column we need to append to our source rows. transcribed. which is what makes this sort possible w/o re-partitioning (which is also quite expensive. granted. instead of buffering the entire Data Set. All rights reserved. Also worth noting here: because we already hashed the data on ItemID and TransactionDate. . Since the aggregator does not need to disturb row-order (for pre-sorted data). stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. even for ‘fast’ disks). transmitted. and we only want to ‘sub-sort’ the data on StoreLocation (this option is only viable for situations where you need to maintain the sort order on the initial keys). this would save a tremendous amount of very expensive disk I/O as sort can hold a few hundred records in memory in most cases (disk I/O is typically several orders of magnitude more costly than memory I/O.

25 44.Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt.75 38.5 3.25 16.75 38.5 27 27 61.5 26. 2006 169 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. as it is much more expensive than non-stable sorts. transcribed.5 33. and it is enabled by default. useful for preserving previous sort orderings).5 61. All rights reserved.25 44.5 3. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17. . Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. transmitted.75 Store Average Item Transaction Amt 2 2 6 6 5 5 26.5 16.5 44 44 1 1 1 1 2 2 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 3 1 5 7 3 7 45 8 9 45 45 78 2 65 23 65 PeekFinalOutput. No part of this publication may be reproduced.25 3.25 44. This sort will look exactly like the other sort stage: Remember to disable the ‘Stable Sort’ option if you do not need it (it will try to maintain row order except as needed to perform the sort.75 38.25 44.5 16. stored in a retrieval system.5 33.25 38.25 16. Output from above solution: Data Set 3: PeekFinalOutput.

5 Store Average Item Transaction 28 28 43 43 6.Information Integration Solutions Center of Excellence Store Location Item ID Transaction Date Transaction Amt 1 1 2 2 1 1 1 1 2 2 1 1 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1 1 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 7 85 7 6 3 98 23 56 534 87 National Average Item Transaction Amt 35. 2006 170 of 179 © 2006 IBM Information Integration Solutions.5 39. we had to sort (on disk) 300.5 180.5 39.25 26. or translated into any language in any form by any means without the written permission of IBM.5 50.000 records. 100. but there is a critical difference. stored in a retrieval system. No part of this publication may be reproduced.5 35.5 23 23 180.5 46 46 6. a single partitioner.25 180. transcribed. There is an even more efficient solution.000 records.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume. All rights reserved.5 6. 000. The second solution only sorts (on disk) 100.5 310.000 records. and two sub-sorts—a much more efficient solution for large data volumes.5 6. It looks very similar to the first solution. transmitted. . a savings of 400.000.000. With the initial solution. Imagine a job with 100 million records as the input.000.5 180.5 35.25 26. and only hashes.5 50.000 records in addition to hashing 300.5 This solution produces the same result but is achieved with only one complete sort.5 35.000.25 26. Parallel Framework Red Book: Data Flow Job Design July 17.5 23 23 26.5 310. That is a LOT of saved processing power.

Comparing the efficiency of this solution with that of number two. transcribed.25 16. except w/o the sort on the output of JoinSourceToAggregator. This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator. . Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. automatic sort insertion. but NOT to use it for hashing. the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed. or translated into any language in any form by any means without the written permission of IBM.5 3. Also. However it will not create additional processes to handle the records and re-order them. All rights reserved. This is a potentially huge savings on large data volumes (remember the previous example).5 27 27 July 17. stored in a retrieval system.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26.5 3. we saved a sub-sort on 100 million records .25 44.Information Integration Solutions Center of Excellence NOTE: In this job. The difference is on DistributeCopiesOfSourceData: Here. for the same reasons. we have chosen to use the StoreLocation column as a part of our sorting key.5 26.a significant savings. 2006 171 of 179 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 1 1 2 2 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 2004/01/01 2004/01/01 1 3 7 3 5 7 8 45 45 9 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Here is the output from this version of the job. Data Set 4: PeekFinalOutput.5 16. No part of this publication may be reproduced. transmitted.5 3. Looks a lot like solution 1.25 16.25 16.25 44. and automatic partition insertion must be turned off vii.

25 44.25 26.5 39. inhibits pipe-lining (by buffering large amounts of data to disk since it needs to see all data before it can determine the resulting sorted sequence)ix.5 180. . or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence 2 2 1 1 2 2 2 2 2 2 2 2 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 78 45 2 65 23 65 44. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 310.75 38.5 35.5 6. 2006 172 of 179 © 2006 IBM Information Integration Solutions.5 Store Average Item Transaction Amt 28 28 43 43 6. All rights reserved. stored in a retrieval system.25 26.5 46 46 6.5 6.5 50. transcribed.5 23 23 23 23 26. by definition.75 38.5 33.5 180. sort. in addition to the heavy penalty paid in disk I/O for using a full sort.5 33.5 50.25 180.5 310.5 180.25 38. transmitted.5 44 44 PeekFinalOutput.25 26. Here is a screen shot of a sort running on 40 million records: Parallel Framework Red Book: Data Flow Job Design July 17.5 35.5 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 23 56 7 85 7 6 3 98 534 87 Finally.75 61.75 38.5 35. No part of this publication may be reproduced.5 39.

instead of waiting until all 40 million records have been sorted (in this instance. transmitted. we are sub-sorting the data we sorted in the previous diagram). you can clearly see that a sub-sort does not inhibit pipe-lining--very nearly the same number of rows have entered and left the sort stage (and NO buffering is required to perform the subsort). or translated into any language in any form by any means without the written permission of IBM. All rights reserved. No part of this publication may be reproduced. no rows have left yet. This situation is analogous to all of the sorts in solution 1. This is because a standard sort requires all rows to be present in order to release the first row. the link sort in solution 2. . Parallel Framework Red Book: Data Flow Job Design July 17. and the link sort in solution three.Information Integration Solutions Center of Excellence As you can see. stored in a retrieval system. This allows down-stream stages to be processing data during the sorting process. requiring a large amount of scratch diskx. 2006 173 of 179 © 2006 IBM Information Integration Solutions. transcribed. although ~5 million records have entered the sort. Here is an example of a ‘sub-sort’: Here.

problem: you created a fileset with 8 nodes. There are other situations where this is valuable but they are much less common. key={ value=TransactionDate }.extra parameters for key. iii ParallelSortMerge Operator Options: -key -. optional.dataset to be resorted/merged. optional -cs -. Join needs to see all of the rows in the current cluster on at least one of the input links. In other cases.descending sort order. 0 or larger. -cs) (mutually exclusive: -asc.property=value pair(s). optional -param -.where null values should sort.report statistics at the end of the run. optional This operator may have following inputs -Sorted -. but not introduce superfluous sorts. you can use the ParallelSortMerge stage to ensure that no matter the degree of parallelism of the writer + reader. This ‘problem’ is addressed by the parallelsortmerge component iii. but the job that reads it only has 4 nodes. Another common problem: You need to Hash and Sort on columns A and C to implement the business logic in one section of the flow. race/gender/ethnicity codes are typical).input field name Sub-options for key: -ci -. without curly braces (mutually exclusive: -ci. 2147483647 or smaller. which would result in having to completely re-sort that dataset despite having a sorted version already. gender codes. default=first -asc or ascending -. However. optional position -. datasets with differing partition counts and you wish to join/merge them.queue length at which to issue a warning.integer.case-insensitive comparison. the sort order will be preserved. less common.case-sensitive comparison. last.specify a key field. key={ value=TransactionDate } })#>eCollectAny op2[3p] (parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt)} ds2: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }. default -nulls -. You could hash only on A. Normally EE would re-partition the data into 4 nodes + destroy your sort order. This would allow you to combine other columns into your hash key to reduce data-skew. A third. presorted. ii A common problem: Suppose you have two (or more). default=10000 -doStats -. exactly one occurrence required <add example here of how psm works> iv Dump score for solution 1 main_program: This step has 16 datasets: ds0: {op0[3p] (parallel SourceData. key={ value=TransactionDate } })#>eCollectAny op4[3p] (parallel APT_TSortOperator(0))} ds4: {op2[3p] (parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) [pp] eSame=>eCollectAny op5[3p] (parallel APT_TSortOperator(1))} ds5: {op3[3p] (parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) [pp] eSame=>eCollectAny op8[3p] (parallel APT_TSortOperator(2))} ds6: {op4[3p] (parallel APT_TSortOperator(0)) . optional records -. but suppose that A has too small a number of unique values (country codes. otherwise a Cartesian product is impossible. 1 or more name -. but in another section you need to hash and sort on columns A and B.resorted dataset. -desc) -warnLevel -. At least one of these dataset must be re-hashed. exactly one occurrence required This operator may have following outputs -reSorted -. optional. optional params -.ascending sort order. key={ value=StoreLocation } })#>eCollectAny op3[3p] (parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt)} ds3: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }.string. value one of first. it’s only true for cases where the key is unique.i This is an over-simplification. default -desc or descending -.DSLink2) eAny=>eCollectAny op1[3p] (parallel DistributeCopiesOfSourceDta)} ds1: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }.

p0] node1[op1.p2] )} op2[3p] {(parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op2.p1] node3[op3.p1] node1[op1.DSLink2) on nodes ( node1[op0. subArgs={ cs } }.p2] )} op3[3p] {(parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op3. subArgs={ cs } } })#>eCollectAny op10[3p] (parallel JoinSourceToAggregator_2.p1] node3[op2.p2] )} op4[3p] {(parallel APT_TSortOperator(0)) on nodes ( node1[op4. key={ value=TransactionDate }.p2] )} op5[3p] {(parallel APT_TSortOperator(1)) on nodes ( node1[op5.p1] node1[op0.p0] node2[op3.p2] )} op6[3p] {(parallel buffer(0)) .p0] node2[op4.p0] node2[op2.p1] node3[op4. key={ value=StoreLocation.DSLink18_Sort)} ds12: {op10[3p] (parallel JoinSourceToAggregator_2.p1] node3[op5.p0] node2[op5.p0] node1[op0.DSLink18_Sort) [pp] eSame=>eCollectAny op11[3p] (parallel buffer(2))} ds13: {op11[3p] (parallel buffer(2)) [pp] eSame=>eCollectAny op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds14: {op12[3p] (parallel buffer(3)) [pp] eSame=>eCollectAny op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds15: {op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2) [pp] eSame=>eCollectAny op14[3p] (parallel PeekFinalOutput)} It has 15 operators: op0[3p] {(parallel SourceData.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op1.[pp] eSame=>eCollectAny op6[3p] (parallel buffer(0))} ds7: {op5[3p] (parallel APT_TSortOperator(1)) [pp] eSame=>eCollectAny op7[3p] (parallel buffer(1))} ds8: {op6[3p] (parallel buffer(0)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds9: {op7[3p] (parallel buffer(1)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds10: {op8[3p] (parallel APT_TSortOperator(2)) [pp] eSame=>eCollectAny op12[3p] (parallel buffer(3))} ds11: {op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1) [pp] eOther(APT_HashPartitioner { key={ value=ItemID.

p0] node2[op9.p1] node3[op13.p2] )} op7[3p] {(parallel buffer(1)) on nodes ( node1[op7. or order.p2] )} op9[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_1) on nodes ( node1[op9.p1] node3[op6. subArgs={ cs } }.e.DSLink2_Sort) [pp] eSame=>eCollectAny op2[3p] (parallel DistributeCopiesOfSourceDta)} ds2: {op2[3p] (parallel DistributeCopiesOfSourceDta) .p2] )} op14[3p] {(parallel PeekFinalOutput) on nodes ( node1[op14.p1] node3[op12.p0] node2[op11.p1] node3[op14.p1] node3[op7.p2] )} It runs 45 processes on 3 nodes.DSLink2_Sort)} ds1: {op1[3p] (parallel DistributeCopiesOfSourceDta.p1] node3[op10. if you move records unnecessarily.DSLink18_Sort) on nodes ( node1[op10. key={ value=TransactionDate } })#>eCollectAny op1[3p] (parallel DistributeCopiesOfSourceDta. your run time will be adversely affected.p1] node3[op11. Since moving records around takes CPU time and extra system calls.p2] )} op13[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_2) on nodes ( node1[op13. a record changes partition.p2] )} op8[3p] {(parallel APT_TSortOperator(2)) on nodes ( node1[op8. vi Dump Score for Solution 2 main_program: This step has 15 datasets: ds0: {op0[3p] (parallel SourceData.p0] node2[op6.p0] node2[op12.p0] node2[op8.p1] node3[op9.p1] node3[op8.p0] node2[op13.p0] node2[op10.p2] )} op12[3p] {(parallel buffer(3)) on nodes ( node1[op12.p0] node2[op7.on nodes ( node1[op6.DSLink2) eOther(APT_HashPartitioner { key={ value=ItemID. v Throughout this document the general meaning of the phrase ‘more efficient’ is fewer record movements--i. fewer times.p2] )} op11[3p] {(parallel buffer(2)) on nodes ( node1[op11.p0] node2[op14.p2] )} op10[3p] {(parallel JoinSourceToAggregator_2.

DSLink2) on nodes ( node1[op0.p0] node2[op3.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta.p0] node1[op0.p2] )} op3[3p] {(parallel SubSortOnStoreLocation) on nodes ( node1[op3.p2] )} op5[3p] {(parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op5.[pp] eSame=>eCollectAny op3[3p] (parallel SubSortOnStoreLocation)} ds3: {op2[3p] (parallel DistributeCopiesOfSourceDta) [pp] eSame=>eCollectAny op4[3p] (parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt)} ds4: {op2[3p] (parallel DistributeCopiesOfSourceDta) [pp] eSame=>eCollectAny op6[3p] (parallel buffer(0))} ds5: {op3[3p] (parallel SubSortOnStoreLocation) [pp] eSame=>eCollectAny op5[3p] (parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt)} ds6: {op4[3p] (parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) [pp] eSame=>eCollectAny op7[3p] (parallel buffer(1))} ds7: {op5[3p] (parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) [pp] eSame=>eCollectAny op8[3p] (parallel buffer(2))} ds8: {op6[3p] (parallel buffer(0)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds9: {op7[3p] (parallel buffer(1)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds10: {op8[3p] (parallel buffer(2)) [pp] eSame=>eCollectAny op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds11: {op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1) [pp] eSame=>eCollectAny op10[3p] (parallel SubSortOnStoreLocation2)} ds12: {op10[3p] (parallel SubSortOnStoreLocation2) [pp] eSame=>eCollectAny op11[3p] (parallel buffer(3))} ds13: {op11[3p] (parallel buffer(3)) [pp] eSame=>eCollectAny op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds14: {op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2) [pp] eSame=>eCollectAny op13[3p] (parallel PeekFinalOutput)} It has 14 operators: op0[3p] {(parallel SourceData.p1] node3[op5.DSLink2_Sort) on nodes ( node1[op1.p1] node1[op0.p1] node3[op4.p2] )} op2[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op2.p0] node2[op1.p2] )} op4[3p] {(parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op4.p0] node2[op4.p1] node3[op2.p1] node3[op1.p1] node3[op3.p0] node2[op5.p0] node2[op2.p2] .

p0] node2[op8. you want auto insertion turned off b/c EE will see that you are ‘missing’ a sort/partitioner and insert one for you. thus introducing the inefficiencies we are trying to avoid.p2] )} op13[3p] {(parallel PeekFinalOutput) on nodes ( node1[op13.p2] )} op10[3p] {(parallel SubSortOnStoreLocation2) on nodes ( node1[op10.p1] node3[op12.p2] )} op7[3p] {(parallel buffer(1)) on nodes ( node1[op7. the distribution of groups across partitions is nearly even for large numbers of groups). vii In this instance.p0] node2[op10.p2] )} op11[3p] {(parallel buffer(3)) on nodes ( node1[op11.p2] )} op9[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_1) on nodes ( node1[op9. Here are the possible outcomes if I hash-partition on ColumnA only: ColumnAColumnBColumnCGroup 1111112113121122123Group 2211212213221222223231232233 .p0] node2[op13.p1] node3[op13.p0] node2[op12.p0] node2[op11. in fact. you could have ALL 5 groups sent to the same partition (this is unlikely.p2] )} It runs 42 processes on 3 nodes.p2] )} op8[3p] {(parallel buffer(2)) on nodes ( node1[op8. viii To understand why this is true. and ColumnB: ColumnAColumnBColumnCGroup 1111112113Group 2121122123Group 3211212213Group 4221222223Group 5231232233 There must be exactly 5 groups identified by the hash algorithm b/c there are exactly 5 unique combinations of ColumnA and ColumnB.p0] node2[op7. This does not mean that these groups will be in unique partitions: consider a job that only has 3 partitions. look at this example: Here is my source data: ColumnAColumnBColumnC111112113121122123211212213221222223231232233 Here are the possible outcomes if I hash-partition on ColumnA.p1] node3[op7. regardless of the number of partitions: if you are running a job with 6 partitions.p1] node3[op11.p1] node3[op9.p1] node3[op10.p2] )} op12[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_2) on nodes ( node1[op12. Any combination of these groups can be in any partition.p1] node3[op8.p0] node2[op9.p1] node3[op6.)} op6[3p] {(parallel buffer(0)) on nodes ( node1[op6.p0] node2[op6. and the likelihood decreases with larger numbers of groups.

In the above example. there are only two groups by hashing on ColumnA only. you will get a very small number of groups. b/c: if all unique values of ColumnA are together. One effect is that if we wanted to aggregate on ColumnA and ColumnB. ix This means that down-stream process will be sitting idle until the sort is completed. larger groups. this grouping is OK. You need to understand your data and make educated decisions about your hashing strategy. It is only true on a per-partition basis. summing ColumnC. x This is a slight oversimplification. even if we ran the job 12-ways. effectively. we want to reduce the number of times that we hash (b/c partitioning costs CPU time). In the scenario that we are discussing in the main document. So hashing on fewer columns resulted in fewer. then all unique combinations of ColumnA and ColumnB are together. . as well as all unique combinations of ColumnA. consuming RAM and process space and offering nothing in return. reduce the parallelism of the job. not for the entire dataset. and ColumnC. we wouldn’t see any improvements in performance over a 2-way job.As you can see. ColumnB. we would have only two groups. which will. We can do this by identifying the intersection of keys needed among all of the hash-partitioners and hashing only on those keys: TransactionDate and ItemId NOTE: if you take this to an extreme. Therefore.

Sign up to vote on this title
UsefulNot useful