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.

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

....................................Information Integration Solutions Center of Excellence 8.......94 9....2 CAPTURING UNMATCHED RECORDS FROM A JOIN.........................2 DEBUGGING ENVIRONMENT VARIABLES......160 Parallel Framework Red Book: Data Flow Job Design July 17........................... 2006 5 of 179 © 2006 IBM Information Integration Solutions........................................87 8.......124 11.....................124 11.129 12..................................3 MINIMIZING RUNTIME PROCESSES AND RESOURCE REQUIREMENTS.......................103 10........................................... JOIN VS............................ All rights reserved...............1 LOOKUP VS......................................................................................................................................129 12....................................................140 APPENDIX B: DATASTAGE NAMING REFERENCE.....................................................94 9...............................................126 11.............................3 THE AGGREGATOR STAGE...................................................................................................................................................119 10.................................................................................................................................................................................................................................................................1 WARNING ON SINGLE-NODE CONFIGURATION FILES....113 10..................................................3 HOW TO ISOLATE AND DEBUG A PARALLEL JOB..........................................................2 MODIFY STAGE...................................................................................................1 DATABASE DEVELOPMENT OVERVIEW.........................................................................................................................................................................3 INFORMIX DATABASE GUIDELINES............................................................................................96 10............................................................... No part of this publication may be reproduced................................................. transcribed..................................................................131 12..................................................................................................................................................................................5 ORACLE DATABASE GUIDELINES..........................................................154 APPENDIX F: SORTING AND HASHING ADVANCED EXAMPLE....................................................................1 TRANSFORMER STAGE............................................124 11...........................................................................................................................................................................................................................................................................127 12 PERFORMANCE TUNING JOB DESIGNS.............................................................................134 APPENDIX A: STANDARD PRACTICES SUMMARY............6 SYBASE ENTERPRISE GUIDELINES.....................................................................................................................................................96 10.......................................................................................2 DB2 GUIDELINES............................................................... or translated into any language in any form by any means without the written permission of IBM.......... stored in a retrieval system.......153 APPENDIX E: ENVIRONMENT VARIABLE REFERENCE...........4 VIEWING THE GENERATED OSH..............95 10 DATABASE STAGE GUIDELINES............ ..............146 APPENDIX C: UNDERSTANDING THE PARALLEL JOB SCORE........................................................................5 INTERPRETING THE PARALLEL JOB SCORE....91 9 COMBINING DATA..............................................4 ODBC ENTERPRISE GUIDELINES.1 HOW TO DESIGN A JOB FOR OPTIMAL PERFORMANCE................................................125 11........................................148 APPENDIX D: ESTIMATING THE SIZE OF A PARALLEL DATA SET..................................120 11 TROUBLESHOOTING AND MONITORING.........................................................................................................................................................................................7 TERADATA DATABASE GUIDELINES..................................4 UNDERSTANDING BUFFERING...........................................2 UNDERSTANDING OPERATOR COMBINATION....................................................................................114 10...94 9.............................................. transmitted...............................................117 10...................................... MERGE....................................................................133 12..................................

2006 6 of 179 After Job © 2006 IBM Information Integration Solutions. Performing row validation to support data quality. transcribed. All rights reserved. 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. stored in a Subroutine retrieval system. transmitted. No part of this publication may be reproduced. and • Provisioning data targets.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. or translated into any language in any form by any means without the written permission of IBM. . Performing transformation from data sources to data targets.

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

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

Parallel Framework Red Book: Data Flow Job Design July 17. transmitted.2. If the entire target table is held in the load-ready Data Set. or translated into any language in any form by any means without the written permission of IBM. .Information Integration Solutions Center of Excellence 1.1 Transformation Jobs In transformation jobs. some of which may be write-through cache Data Sets. stored in a retrieval system. 2006 9 of 179 © 2006 IBM Information Integration Solutions. 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. All rights reserved. transcribed. 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. that Data Set qualifies as write-through cache and may be used as source data instead of the target table. No part of this publication may be reproduced.

transcribed. stored in a retrieval system. No part of this publication may be reproduced.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. 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. 2006 10 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.

stored in a retrieval system. the cache is deleted. transmitted. 2006 11 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence 1. No part of this publication may be reproduced. 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. If the key column is NULL. The non-validated records.2 Hybrid Jobs The following example hybrid job demonstrates several interesting techniques that might be used in more complex jobs. or translated into any language in any form by any means without the written permission of IBM. The de-duplicated records are re-grouped and ordered before calculation of the terminating keys. The merged records are grouped and ordered before being de-duplicated to remove obsolete records. transcribed. we replicated the Oracle structure (lower input) for each country found in the write-through cache country dataset (upper input). producing an ordered and linked associative table.2. and the write-through cache records from the last load of the target database are merged. it is rejected by the transformer to a reject port and the validation is not performed for those records. if the load fails. . Some of the more interesting solutions in this job are circled. All rights reserved. forcing other jobs that might depend on this data to access the existing (not updated) target database table. the validated records. This job also loads the target database table and creates write-through cache. In this case. By specifying a full-outer join we produce a Cartesian product dataset. The key column for a Referential Integrity check is validated by a Transformer stage. 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). Parallel Framework Red Book: Data Flow Job Design July 17. In this case.

3 Provisioning Jobs This example provisioning job demonstrates the straightforward approach to simple provisioning tasks. All rights reserved. transcribed.Information Integration Solutions Center of Excellence 1. 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. stored in a retrieval system. transmitted. 2006 12 of 179 © 2006 IBM Information Integration Solutions.2. .

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

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

and production) as appropriate. system integration. /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. . error and reject files.Information Integration Solutions Center of Excellence /Staging /dev /si /qa /prod /Project_A /Project_A /Project_A /Project_A /archive /archive /archive /archive . location of source data files... 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. target data files. /Project_Z July 17.. Within each deployment directory. /Staging /dev Top-Level Directory development data tree. stored in a retrieval system.. files are separated by Project name as shown below. data directories are implemented for each deployment phase of a job (development. 2006 15 of 179 . transcribed. /Project_Z Within the separate Staging file system.. . All rights reserved. If the file system is not /archive /archive /archive shared across multiple servers.. No part of this publication may be reproduced./archive of these development phases may be present on a local file not all system. transmitted. qa..

. transmitted. or translated into any language in any form by any means without the written permission of IBM. Environment variables are a critical portability tool.Information Integration Solutions Center of Excellence 2.2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities. A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project.. 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. stored in a No part of this /Project_Z /Project_Z retrieval system./Project_Z publication may be reproduced.. through the use of environment variables. The directory structure will be made transparent to the DataStage application. or it could require scripts for example integrating with an Enterprise Scheduler.. 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. 2006 .. 16 of 179 . .. /Project_Z . All rights reserved...1. transcribed. which will enable DataStage applications to move through the life cycle without any code changes. such as the operating system. another application or middle ware. The integration can be as simple as a file system for housing source files. July 17. This will provide a convenient vehicle to group and manage resources used by a project..

Parallel Framework Red Book: Data Flow Job Design July 17.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 (. not all of these development phases may be present on a local file system. 2006 17 of 179 © 2006 IBM Information Integration Solutions. transcribed. stored in a retrieval system. 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. a copy of dsenv and copies of DSParams. utilities. transmitted.$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. qa. No part of this publication may be reproduced. 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. DataStage routines. 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. BuildOps. or translated into any language in any form by any means without the written permission of IBM. If the file system is not shared across multiple servers. system integration. All rights reserved. and production) as appropriate. directories are created for each deployment phase of a job (development.

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

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

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

Therefore Directory names should be Word Capitalized and separated by either an underscore or a space.2.4. . No part of this publication may be reproduced.Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage. Below this level. However with the limit of 18 characters the name is most often composed of abbreviations. or translated into any language in any form by any means without the written permission of IBM. it can contain alpha-numeric characters and it can contain underscores. DataStage enforces the top level Directory Structure for different types of Objects (for example. Table definitions…). are Alpha Numeric and can also contain both Spaces and Underscores. which can provide impact analysis. transmitted. 2.3 Job Category Naming Within Designer. and Production (“Prod”). For example. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17. developers have the flexibility to define their own Directory or Category hierarchy.4. 2. All rights reserved. The name of a DataStage Project may only be 18 characters in length.2. Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”).1 Projects Each DataStage Project is a standalone repository. Routines.4. It may or may not have a one to one relationship with an organizations’ project of work. Jobs. stored in a retrieval system. as well as documentation and auditing. The suffix of a Project name should be used to identify Development (“Dev”).2. Category Names can be long. allowing related objects to be grouped together. Test (“Test”).2. This factor often can cause terminology issues especially in teamwork where both business and developers are involved.2 Category Hierarchy DataStage organizes objects in its repository by Categories. 2006 21 of 179 © 2006 IBM Information Integration Solutions. transcribed.4 Naming Conventions by Object Type 2. Shared Containers. dialog box fields that specify a new category have only one input area for defining the Category name. 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.

.Information Integration Solutions Center of Excellence Figure 5: Creating Category Hierarchies The main reason for having Categories is to group related objects. All rights reserved. transmitted. 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. 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. Categorization by Functional Module For a given application. with sub-levels for individual functional modules. a Category level should only contain objects that are directly related. or translated into any language in any form by any means without the written permission of IBM. not within a category. Within each functional module category. For example. Jobs and Job Sequences are grouped together in the same scope as the technical design documents. For example. No part of this publication may be reproduced. Where possible. 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. transcribed. stored in a retrieval system. all Jobs and Job Sequences will be grouped in a single parent Category. 2006 22 of 179 © 2006 IBM Information Integration Solutions.

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

4. An alternative implementation is to set the “Data source name” to that of the source system or schema. e. transcribed.: Datasets. In this example. Parallel Framework Red Book: Data Flow Job Design July 17.2 Category Hierarchy.2. 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.g. with sub-categories intended to identify the type of the source. the TableDefs have been grouped into a master category of Custom. and underscores only. Each subject area will have a master category. 2006 24 of 179 © 2006 IBM Information Integration Solutions. numbers. Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”.Information Integration Solutions Center of Excellence should be deleted as soon as possible.2. Because the name of can be long. e. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”. This will be discussed further in Section 2. . Job and Job Sequence names should be descriptive and should use word capitalization to make them readable. stored in a retrieval system.g: DWPH1 or ECRP. Figure 7: Table Definition Categories 2. transmitted. All rights reserved. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters. 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.4.

When a Shared Container is used.4. stored in a retrieval system. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. No part of this publication may be reproduced. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2. 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.8 Links Within a DataStage Job.2.Information Integration Solutions Center of Excellence 2.2. 2. Furthermore. links are objects that represent the flow of data from one stage to the next.7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores. transmitted. All rights reserved. 2006 25 of 179 © 2006 IBM Information Integration Solutions. stage editors identify links by name. or translated into any language in any form by any means without the written permission of IBM. The class word suffix is “Parm”.4. having a descriptive link name reduces the chance for errors (for example. instead of using the default “DSLink#” (where “#” is an assigned number). Therefore the parameter name must be made readable using Capitalized words separated by underscores. It is optional as to whether you decide to change this code to something meaningful. To differentiate between Parallel Shared Containers and Server Shared Containers. links represent the flow of a message from one activity / step to the next. 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. It is particularly important to establish a consistent naming convention for link names.6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. a character code is automatically added to that instance of its use throughout the project. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17. Within the graphical Designer environment.4. so word capitalization should be used for readability. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. Within a Job Sequence. transcribed.2. during Link Ordering). .

transmitted. stage names must be unique. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12.4. 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. or 4 character abbreviation should be used for the Class Word suffix. All rights reserved. based on the order the object was added to the flow. reject) to reinforce the visual cues of the Designer canvas: o “Ref” for reference links (Lookup) o “Rej” for reject links (Lookup.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. Transformer. Sequential File. etc) • The type of movement may optionally be part of the Class Word. 3. 2006 26 of 179 © 2006 IBM Information Integration Solutions. after the subject name and subject modifier. . through Job reporting). 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.) • 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. a 2. Within a Job or Job Sequence. always specify a unique link name within a particular Job or Job Sequence by including a number. (The DataStage Designer does not require link names on different stages to be unique. These names are based on the type of stage (object) and a unique number. the same name may be appropriate for multiple links. stored in a retrieval system.4. Database. the link name should include the link type (reference.2. Merge.2 DataStage Naming Reference. 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. In this case. No part of this publication may be reproduced. Instead of using the full object name. transcribed.Information Integration Solutions Center of Excellence example. establishing standardized link names makes it easier to understand results and audit counts.

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.

All rights reserved. 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. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. transmitted. Oracle Enterprise stage read: Read the GLOBAL. These descriptions will appear in the job documentation automatically generated from jobs and sequencers adhering to the standards in this document. More complex operators or operations should have correspondingly longer and more complex explanations on this tab. ODBC Enterprise stage read: Read the GLO. 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. 2006 30 of 179 © 2006 IBM Information Integration Solutions.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. There are no selection criteria in the WHERE clause.RcR_GLOBAL_BUSN_CAT_TYP table from jpORACLE_SERVER using the ODBC driver. stored in a retrieval system. There are no selection criteria in the WHERE clause. . Each stage should have a short description of its function specified within the stage properties.GLOBAL_REST_CHAR table from jpORACLE_SERVER using the Oracle Enterprise operator.

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

stored in a retrieval system. These files can then be checked into the external source code control system. Unless there is a need to parse information in the export file. user-defined routines. table definitions. and so on.for example. although the XML file is generally much larger. Typically. DataStage does not directly integrate with source code control systems. when a set of objects has completed initial development.DSX is the recommended export format.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). transcribed. Rather. unit test.XML. It is the responsibility of the DataStage developer to maintain DataStage objects within the source code system. Designer facilitates iterative job design. Grouping related objects by folder can simplify the process of exchanging information with the external source code control system. It would be cumbersome to require the developer to check-in every change to a DataStage object in the external source code control system. Parallel Framework Red Book: Data Flow Job Design July 17.) from the repository as clear-text format files. it is important that an identified individual maintains backup copies of the important job designs using . custom stage types. etc. operating system backups of the project repository files only establish a “point in time”. PVCS. .4.DSX file exports to a local or (preferably) shared file system. 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. No part of this publication may be reproduced. . they would not be an effective backup strategy. transmitted. but it does offer the ability to exchange information with these systems. organized into specific releases for version control. 2. b) Define rules for exchange with source code control As a graphical development environment. you can export objects (job designs. The export file format for DataStage objects can be either . SCCS) are useful for managing the development lifecycle of all components of an application. The Manager client is the primary interface to the DataStage object repository. Furthermore. All rights reserved. 2006 32 of 179 © 2006 IBM Information Integration Solutions. Both formats contain the same information. and cannot be used to restore individual objects. rules should be defined for when this transfer should take place. Using Manager.DSX (DataStage eXport format) or . milestone points on the development lifecycle are a good point for transferring objects to the source code control system . For these reasons.Information Integration Solutions Center of Excellence Source code control systems (such as ClearCase. This object grouping also helps establish a manageable “middle ground” between an entire project exports and individual object exports.

• Select the object or category in the Manager browser. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu. transmitted.2 Using Object Categorization Standards As discussed in Section 2. or by the individual DataStage developer.2 Category Hierarchy.4. The DataStage Manager can export at the Project. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.4.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process. stored in a retrieval system. 2. transcribed. . the developer should create a local backup prior to implementing any extensive changes. These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide. establishing and following a consistent naming and categorization standard is essential to the change management process. No part of this publication may be reproduced.4.2. There is no server-side project export facility. All exports from the DataStage repository are performed on the Windows workstation. It can be done interactively by the developer or project manager using the Manager client. In either case. All rights reserved. July 17. 2. and individual Object levels. The DataStage client includes Windows command-line utilities for automating the export process. Assigning related objects to the same category provides a balanced level of granularity when exporting and importing objects with external source code control systems. or translated into any language in any form by any means without the written permission of IBM. as explained in this section. Category.Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group.

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. The filename for export is specified in the “Export to file:” field at the top of the Export dialog. check-in the exported . Make sure all objects are saved and closed before exporting. the option “Selection: By category” should be specified in the “Options” tab. Figure 11: Manager Export Options To export a group of objects to a single export file. transmitted. . All rights reserved. transcribed. No part of this publication may be reproduced. 2006 34 of 179 © 2006 IBM Information Integration Solutions. the import of objects from an external source code control system is a straightforward process. If you wish to include compiled Transformer objects for a selected job.DSX file 2. or automated through command-line utilities.4 Import from Source Code Control System In a similar manner. • Using your source code control utilities. stored in a retrieval system.4. make sure the “Job Executables” category is checked.Information Integration Solutions Center of Excellence NOTE: Objects cannot be exported from DataStage if they are open in Designer. Parallel Framework Red Book: Data Flow Job Design July 17.

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. or translated into any language in any form by any means without the written permission of IBM.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. dsrpcd is started when the server installed. command-line import utilities are available for both Windows workstation and DataStage server platforms. No part of this publication may be reproduced. • • Use the source code control system to check-out (or export) the . or using the Multi-Job Compile tool.DSX file will place the object in the same DataStage category it originated from. as documented in the DataStage Development Kit chapter of the Parallel Job Advanced Developer’s Guide. then compile the imported objects from Designer. For test and production environments.DSX file using Manager. The Windows workstation utilities (dsimport and dscmdimport) are documented in the DataStage Manager Guide. After selecting your file. 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. By default (in a root installation). transmitted.DSX file to your client workstation. Although operating system environment variables can be set in multiple places. This means that if necessary it will create the Job Category if it doesn't already exits. 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. 2.Information Integration Solutions Center of Excellence Unlike the export process. 2006 35 of 179 © 2006 IBM Information Integration Solutions. If the objects were not exported with the “Job Executables”. Choose “Import DataStage Components” from the “Import” menu. stored in a retrieval system. . and should start Parallel Framework Red Book: Data Flow Job Design July 17. it is possible to import the job executables from the DataStage server host using the dsjob command-line. Import objects in the . All rights reserved. Figure 12: Manager Import options • • The import of the . transcribed.

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

.5.2 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. or translated into any language in any form by any means without the written permission of IBM. transcribed.1. processes.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. 2. This setting should be disabled by default.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment. These settings can be made at the project level.4. 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. It may be helpful to create a Job Template and include these environment variables in the parameter settings. The count is per operator per partition. $UNSET 2. but part of every job design so that it can be easily enabled for debugging purposes. This variable should be included in all job parameters so that it can be easily changed at runtime. and Data Sets. (see 12. Extremely useful for understanding how a job actually ran in the environment. Outputs actual runtime metadata (schema) to DataStage job log. This setting should be disabled by default. 2006 37 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. providing detailed information about actual job flow including operators. transmitted.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. APT_SORT_INSERTION_CHECK_ONLY). 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. Does not report PIDs of DataStage “phantom” processes started by Server shared containers. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. stored in a retrieval system. No part of this publication may be reproduced. Several environment variables are evaluated only for their presence in the environment (for example. or may be set on an individual basis within the properties for each job. but part of every job design so that it can be easily enabled for debugging purposes. This section is intended to call attention to some specific environment variables.5. and to document a few that are not part of the documentation. Outputs EE score dump to the DataStage job log.

1A platforms only: On Tru64 platforms. Parallel Framework Red Book: Data Flow Job Design July 17.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). a number of environment variables will be mentioned for tuning the performance of a particular job flow. assisting in debugging. you must define the environment variable $APT_IO_NOMAP On Tru64 5. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. No part of this publication may be reproduced.1. the environment variable $APT_PM_NO_SHARED_MEMORY should be set to 1 to work around a performance issue with shared memory MMAP operations.5. stored in a retrieval system.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. 2006 38 of 179 © 2006 IBM Information Integration Solutions. or changing the default behavior of specific Enterprise Edition stages. . All rights reserved. The environment variables mentioned in this document are summarized in Appendix D: Environment Variable Reference. 2. This setting instructs EE to use named pipes rather than shared memory for local data transport. transmitted. transcribed.

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

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.

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

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

each must be done in its own Lookup. Send the reject stream to an *. 2006 45 of 179 © 2006 IBM Information Integration Solutions. Pass successful lookups to the output stream.rej extension is used when rejects require investigation after a job run. All rights reserved. If a file is created by this option. it must have a *.Information Integration Solutions Center of Excellence 3. .7. The *. Rejects should not exist and should stop the job. Rejects should not exist but should not stop the job.err file or tag and merge with the output stream. to enforce error management ONLY ONE REFERENCE LINK is allowed on a Lookup stage. Alternatively. transcribed. Fail Abort job on lookup failure Reject Reject lookup failures to the reject stream. however. Drop Drop lookup failures from the input stream. stored in a retrieval system. 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. transmitted.rej or *. and rejects can occur if the key fields are not found in the reference data. a local error handler based on a shared container can be used. 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. No part of this publication may be reproduced. Lookup Stage Option Drop if lookup fields are necessary down stream or Continue if lookup fields are optional Send the reject stream to an *. Pass successful lookups to the output stream. 2 3 4 3.3 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. they only need to be recorded but not acted on.2 Reject Handling with the Lookup Stage The Lookup stage compares a single input stream to one or more reference streams using keys. The reject option should be used in all cases where active management of the rejects is required. Use the Fail option.rej file or tag and merge with the output stream. Furthermore.err file extension. the *.7. Pass successful lookups to the output stream.err extension when rejects can be ignored but need to be recorded. If there are multiple validations to perform. or translated into any language in any form by any means without the written permission of IBM. and should be reviewed by the Data Steward.

however. stored in a retrieval system.4 Reject Handling with target database stages Some database stages (such as DB2/UDB Enterprise. Rejects can exist in the data. If a file is created by this option. Send the reject stream to a reject file and halt the job.Information Integration Solutions Center of Excellence 1. transmitted. transcribed.rej extension is used when rejects require investigation after a job run. Pass rows that fail to be written to the reject stream. Rejects should not exist and should stop the job.7. Satisfies requirements for a reject conditional output stream. they only need to be recorded but not acted on.err file or tag and merge with the output stream. A message is always written to the Director log which details the count of rows successfully read and rows rejected. 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. ODBC Enterprise. the *.rej file or tag and merge with the output stream. and Oracle Enterprise) offer an optional reject link that can be used to capture rows that cannot be written to the target database. Send the reject stream to an *. it must have a *. and be reviewed by the Data Steward. OR 2.err extension when rejects can be ignored but need to be recorded. Alternatively.rej or *. The reject option should be used in all cases where active management of the rejects is required. Otherwise. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. To capture rejects from a target database. a shared container error handler can be used. Transformer Stage Option Funnel the reject stream back to the output stream(s). . 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. 2006 46 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. Rejects should not exist but should not stop the job. Cannot satisfy requirements of any conditional output stream and is rejected by the default output stream. The *. a shared container error handler is used. Alternatively. 3. Send the reject stream to an *. If a file is created from the reject stream.err file extension. a reject link must exist on that stage.rej file extension. reject rows will not be captured. it must have a *. 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.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 Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job. 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. 4. 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. 3. stored in a retrieval system.rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504. one of “rej” or “err”. Reject link exists. 2 Rejects should not exist but should not stop the job.”. .7.5.rej file. and should be reviewed by the Data Steward. The job name (jpJOB_NAME) and a underscore “_”.Rejects are tracked by count only.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: 1.7. transcribed. This section deals with both methods of handling errors.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.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 5. 2. For example. 3. The appropriate file type. transmitted. All rights reserved. and gathered using a Funnel stage that feeds a Sequential File stage. Send the reject stream to a *. 3. The project name (jpPROJECT_NAME) and a underscore “_”. The project phase (jpENVIRON) and a underscore “_”. Only records that match the given table definition and database constraints are written. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The job serial number (jpJOBSERIALNO) and a period “.

2006 48 of 179 © 2006 IBM Information Integration Solutions. . 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. transcribed.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. The Transformer stage adds the required key columns. transmitted. No part of this publication may be reproduced. Figure 13: Error Processing Components Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. stored in a retrieval system.

stored in a retrieval system. 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. 2006 49 of 179 © 2006 IBM Information Integration Solutions. transcribed. transmitted. 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. Track*) to a single output column.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.

7. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream. Connect the reject port to a Transformer stage where columns are set to specific values. or translated into any language in any form by any means without the written permission of IBM. Switch Transformer In this example. OR with columns contain illegal values for some operation performed on said columns. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values.Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3. stored in a retrieval system. 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. 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.5. 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. . In either case. Connect the reject port to a Transformer stage where columns are set to specific values. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key.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. A failed switch will reject an intact input row show key fails to resolve to one of the Switch output stream. One or more columns may have been selected for replacement when a reference key is found. All rights reserved. transcribed. 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. The following table identifies the tagging method to be used for the previously cited operators. Connect the output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams. transmitted. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. 2006 50 of 179 © 2006 IBM Information Integration Solutions.

Server Edition components limit overall performance of large-volume job flows since many components such as the BASIC Transformer use interpreted psuedo-code. 2006 51 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 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 17.8. severely impacting scalability and network resources. All rights reserved.8 Component Usage DataStage Enterprise Edition offers a wealth of component types for building ETL flows. No part of this publication may be reproduced. transcribed. 3.1 Server Edition Components Avoid the use of Server Edition components in parallel job flows. . 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. transmitted. In clustered an MPP environments Server Edition components only run on the primary (conductor) node.Information Integration Solutions Center of Excellence Figure 16: Error Processing Lookup example 3. This section provides guidelines appropriate use of various stages when building a parallel job flows.

and Production Automation.Renaming Columns .2 Copy Stage For complex data flows.Server shared containers Note that BASIC Routines are still appropriate. Since the Copy stage does not require an output link.8. transmitted. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. For this reason. 3. transcribed. 3.8. it is important to minimize the number of transformers. and to use other stages (such as Copy) when derivations are not needed.Job Design placeholder between stages (unless the Force option =true. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run. . • The Copy stage should be used instead of a Transformer for simple operations including: . 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Used in this manner.8. and necessary. it is best to develop a job iteratively using the Copy stage as a “placeholder”. NOTE: Because parallel Data Sets are platform and configuration-specific. All rights reserved. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs.4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. o Unless the Force property is set to “True”.Dropping Columns July 17. Enterprise Edition will optimize this out at runtime) . Data Sets offer parallel I/O on read and write operations. it can be used at the end of a data flow o For simple jobs with only two stages. This is discussed in more detail in Parallel Framework Standard Practices: Administration.BASIC Routines . Management.BASIC Transformers . in sort order. No part of this publication may be reproduced. and in Enterprise Edition native format. they should not be used for long-term archive of source data.Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: . the Copy stage should be used as a placeholder so that new stages can be inserted easily should future requirements change. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime. without overhead for format or data type conversions. 3. always write to parallel Data Sets. Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form.3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs.

if possible. stored in a retrieval system. • NEVER use the “BASIC Transformer” stage in large-volume job flows.5. • Because the parallel Transformer is compiled. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines. No part of this publication may be reproduced. Instead.2: Default and Explicit Type Conversions] Note that rename.2: Modify Stage. performance should this should be tested in isolation to identify specific cause of bottlenecks. drop (if Runtime Column Propagation is disabled). implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. transcribed.Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4. The only time that Filter or Switch should be used is when the selection clauses need to be parameterized at runtime. null handling. Even then its use should be restricted and the routines should be converted as soon as possible.7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B. Consider. . user-defined functions and routines can expand parallel Transformer capabilities.1.1. and character string trimming.3 Then B=”X” If A=4. it is faster than the interpreted Filter and Switch stages. See Section 8. 3.5 BuildOp stages BuildOps should only be used when: . All rights reserved. • The Modify stage can be used for non-default type conversions. and default type conversion can also be performed by the output mapping tab of any stage. or translated into any language in any form by any means without the written permission of IBM.6. Parallel Framework Red Book: Data Flow Job Design July 17.8. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible.2. 2006 53 of 179 © 2006 IBM Information Integration Solutions. the derivation expression: If A=0.Existing Transformers do not meet performance requirements As always.Complex reusable logic cannot be implemented using the Transformer or . For example. transmitted.

Numeric Float. uint8 int16. LongVarChar. stored in a retrieval system. day. transmitted. Each SQL data type maps to an underlying data type in the Enterprise Edition engine. VarChar NChar. . Real Double TinyInt SmallInt Integer BigInt1 Binary. All rights reserved. (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. 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. with resolution to seconds Time of day. compatible with IBM packed decimal format. and year Packed decimal.Information Integration Solutions Center of Excellence 4 DataStage Data Types The DataStage Designer and Manager represent column data types using SQL notation. VarBinary Unknown. uint16 int32. 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. unit32 int64. transcribed. Bit. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 54 of 179 © 2006 IBM Information Integration Solutions. 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. LongVarBinary. LongVarChar. VarChar Time Time Timestamp Timestamp date decimal sfloat dfloat int8. NVarChar. Single field containing both date and time value with resolution to microseconds. with resolution of microseconds (Specify microseconds Extended option) Single field containing both date and time value with resolution to seconds. The following table summarizes the underlying data types of DataStage Enterprise Edition: SQL Type Internal Type Size Description Date Decimal. Char. LongNVarChar Char. 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. 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.

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. A blank cell indicates that no conversion is provided.1. or translated into any language in any form by any means without the written permission of IBM. 4. No part of this publication may be reproduced. e = You can use a Modify or a Transformer conversion function to explicitly convert from the source field type to the destination field type. and extended Char. (They are specified as such by selecting the Extended check box for the column in the Edit Meta Data dialog box. 2006 © 2006 IBM Information Integration Solutions. VarChar. 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.) An Extended field appears in the columns grid. Default type conversions take place across the stage output mappings of any Enterprise Edition stage.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.Information Integration Solutions Center of Excellence 4. however. String data represents unmapped bytes. The NChar. stored in a retrieval system. All rights reserved. VarChar. NVarChar. specify that these data types are extended. or LongVarChar columns have ‘Unicode’ in this field. and LongNVarChar types relate to underlying ustring types so do not need to be explicitly extended. The Char. parallel jobs support two types of underlying character data types: strings and ustrings. ustring data represents full Unicode (UTF-16) data. 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. transcribed. in which case they are taken as ustrings and do require mapping.1 Strings and Ustrings If NLS is enabled on your DataStage server. .1. You can. transmitted.

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

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). the source value propagates. transmitted. 2006 57 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. out-of-band null representation for NULL values. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. All rights reserved. destination value is never null.1. Source value propagates. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. Source value or null propagates. 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. . July 17. When reading from Data Set and database sources with nullable columns. care must be taken to avoid data rejects.Information Integration Solutions Center of Excellence IMPORTANT: When processing nullable columns in a Transformer stage. No part of this publication may be reproduced. transcribed.1: Transformer NULL Handling and Reject Link. See Section 8. Source Field not Nullable Nullable not Nullable Nullable Destination Field not Nullable Nullable Nullable not Nullable Result Source value propagates to destination. When reading from or writing to Sequential Files or File Sets. Enterprise Edition uses the internal. If the source value is not null.

Information Integration Solutions Center of Excellence If the source value is null. a fatal error occurs. as long as each stage in the shared container has RCP enabled on their stage Output properties. 4.3 Runtime Column Propagation Runtime column propagation (“RCP”) allows job designs to accommodate additional columns beyond those defined by the job developer. or translated into any language in any form by any means without the written permission of IBM. for example the Sequential File stage. All rights reserved. transcribed. No part of this publication may be reproduced. rather than using a large number of jobs with hard-coded table definitions to perform the same tasks. 2006 58 of 179 © 2006 IBM Information Integration Solutions. . Using RCP. Furthermore. Parallel Framework Red Book: Data Flow Job Design July 17. Before a DataStage developer can use RCP. the remaining columns pass through at runtime. Using RCP judiciously in a job design facilitates re-usable job designs based on input metadata. RCP facilitates re-use through parallel shared containers. Some stages. it must be enabled at the project level through the Administrator client. transmitted. stored in a retrieval system. only the columns explicitly referenced within the shared container logic need to be defined. allow their runtime schema to be parameterized further extending re-use through RCP.

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

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

if the Preserve Partitioning flag was set. stored in a retrieval system. In these instances.1. There are some cases when the input stage requirements prevent partitioning from being preserved. Each partition receives the entire Data Set.Information Integration Solutions Center of Excellence some stages (Sort.2. 5.1. Distributes rows evenly across partitions. in a round robin partition assignment. a warning will be placed in the Director log indicating that Enterprise Edition was unable to preserve partitioning for a specified stage. and its state is stored in persistent Data Sets. All rights reserved. as shown in on the right: Figure 22: Preserve Partitioning option The Preserve Partitioning flag is part of the Data Set structure. 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.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. . the degree of parallelism remains unchanged: 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. Instead. 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. although it can be explicitly set or cleared in the “Advanced” stage properties of a given stage. transmitted. 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. between servers). a parallel Sort). in the case of a cluster or Grid. as illustrated on the right: Same partitioning doesn’t move data between partitions (or. Because Same does not redistribute existing partitions. for example). it retains the partitioning from the output of the upstream stage.1 Same Partitioning Same partitioning in fact performs no partitioning to the input Data Set. transcribed. 5. 2006 61 of 179 © 2006 IBM Information Integration Solutions.

similar to dealing cards: Round robin partitioning has a fairly low overhead. regardless of the current $APT_CONFIG_FILE 5. transcribed. 3 2 1 0 . it is rarely used in real-world data flows. 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). On clustered and Grid implementations.2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment. All rights reserved. . Random partitioning has a slightly higher overhead than Round Robin partitioning. the order that rows are assigned to a particular partition will differ between job runs. the downstream stage runs with the degree of parallelism used to create the Data Set. . Since the random partition number must be calculated. 2006 62 of 179 © 2006 IBM Information Integration Solutions. Entire partitioning may have a performance impact. As a result. Random partitioning evenly distributes rows across partitions. . but using a random assignment. stored in a retrieval system.2. …8 7 6 5 4 3 2 1 0 ENTIRE . or translated into any language in any form by any means without the written permission of IBM. as the complete Data Set must be distributed across the network to each node. Since optimal parallel processing occurs when all partitions have the same workload. 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). 5.1.2.1. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5. 3 2 1 0 . Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning. No part of this publication may be reproduced. While in theory Random partitioning is not subject to regular data patterns that might exist in the source data.4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially. .1.2.3 Random Partitioning Like Round Robin. transmitted.

No part of this publication may be reproduced. and there are a large number of unique values. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. If the source data values are evenly distributed within these key column(s).3.1. 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. 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. ensuring that records with the same values in those key column(s) are assigned to the same partition. Join) require processing on groups of related records.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. or translated into any language in any form by any means without the written permission of IBM. Keyed partitioning is used when business rules (for example. All rights reserved. then the resulting partitions will be of relatively equal size. transcribed. transmitted. stored in a retrieval system.3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns.Information Integration Solutions Center of Excellence 5. 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. . 2006 63 of 179 © 2006 IBM Information Integration Solutions. As an example of hashing. Remove Duplicates) or stage requirements (for example. For DB2 Enterprise Server Edition with DPF (DB2/UDB) only – matches the internal partitioning of the specified source or target table. 5.

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. No part of this publication may be reproduced. transmitted.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. stored in a retrieval system. All rights reserved. .

there are more instances of “Ford” than “Dodge”. individual key column values have no significance for partition assignment. The remainder is used to assign the value to a given partition: partition = MOD (key_value / number of partitions) Like hash. 2006 65 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 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. Also note that only the unique combination of key column values appear in the same partition when used for hash partitioning. the partition size of modulus partitioning will be equally distributed as long as the data values in the key column are equally distributed.3. 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. regardless of the actual number of nodes in the parallel configuration file.2 Modulus Partitioning Modulus partitioning uses a simplified algorithm for assigning related records based on a single integer key column. Using the same source Data Set. . transcribed. When using hash partitioning on a composite key (more than one key column). Also note that in this example the number of unique values will limit the degree of parallelism. It performs a modulus operation on the data value using the number of partitions as the divisor. No part of this publication may be reproduced.1. which would impact performance.Information Integration Solutions Center of Excellence In this case. transmitted. the key column combination of LName and FName yields improved data distribution and a greater degree of parallelism. 5. All rights reserved.

3. Also. partition balance will be skewed. A Range Map file is specific to a given parallel configuration file. To achieve this balanced distribution.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”). or for a non-integer key column. these rows will be assigned to either the first or the last partition. if the incoming Data Set is sequential and ordered on the key column(s). In these instances. when writing data to a target DB2 database using the DB2/UDB Enterprise stage. transmitted. by default. . Range partitioning ensures balanced workload by assigning an approximately equal number of rows to each partition. No part of this publication may be reproduced. And. unlike Hash and Modulus partitioning where partition skew is dependent on the actual data distribution. the Range Map file can be re-used. stored in a retrieval system.Information Integration Solutions Center of Excellence Since modulus partitioning is simpler and faster than hash.3.3 Range Partitioning As a keyed partitioning method. typically where the incoming data values and distribution are consistent over time. 2006 66 of 179 © 2006 IBM Information Integration Solutions. 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. it should be used if you have a single integer key column. if new data values are processed outside of the range of a given Range Map. Range partitioning assigns rows with the same values in one or more key column(s) to the same partition. DB2 partitioning can only be specified for target DB2/UDB Enterprise stages. and the second to actually partition the data within a flow using the Range Map. Parallel Framework Red Book: Data Flow Job Design July 17. defeating the intention of Range partitioning. In another scenario to avoid. To maintain partitioning on data read from a DB2/UDB Enterprise stage. Range partitioning will result in sequential processing. Range partitioning must read the Data Set twice: once to create a Range Map file. transcribed. All rights reserved. 5. data is partitioned to match the internal partitioning of the target DB2 table using the DB2 partitioning method.1. use Same partitioning on the input to downstream stages. or translated into any language in any form by any means without the written permission of IBM. depending on the value.1. Modulus partitioning cannot be used for composite keys. 5. Using the DB2/UDB Enterprise stage. Given a sufficient number of unique values. It is important to note that if the data distribution changes without recreating the Range Map. data is read in parallel from each DB2 node.

right-click anywhere in the window. starting at zero. All rights reserved. No part of this publication may be reproduced. in some cases. and each stage is processing an equal number (12. if specified.2Understanding the Parallel Job Score. and select the “Show Instances” option. as illustrated below.Information Integration Solutions Center of Excellence 5.500) of rows for an optimal balanced workload. The partition number is appended to the stage name for messages written to the Director log. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True. This is very useful in determining the distribution across parallel partitions (skew).4. . In this instance. 2006 67 of 179 © 2006 IBM Information Integration Solutions. 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. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name). transmitted.2 Monitoring Partitions At runtime. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. Partitions are assigned numbers. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. databases) c) and. 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. stored in a retrieval system. Specific details on interpreting the parallel job score can be found in 12. transcribed. or translated into any language in any form by any means without the written permission of IBM.

minimizing overall run time. . Parallel Framework Red Book: Data Flow Job Design July 17. 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. 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. transmitted. stored in a retrieval system. as shown below: Figure 27: Data Set Management Tool In a non-graphical way. 2006 68 of 179 © 2006 IBM Information Integration Solutions.3 Partition Methodology Given the numerous options for keyless and keyed partitioning. while minimizing overhead. the “Data Set Management” tool (available in the Tools menu of Designer. 5. Director. transcribed. or translated into any language in any form by any means without the written permission of IBM. the orchadmin command line utility on the DataStage server can also be used to examine a given parallel Data Set. All rights reserved. or Manager) can be used to identify the degree of parallelism and number of rows per partition for an existing persistent Data Set. No part of this publication may be reproduced.

. Change Apply. 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. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). Using the above objectives as a guide. Join. transmitted. persistent Data Sets can be used to retain the partitioning and sort order. 2006 69 of 179 © 2006 IBM Information Integration Solutions. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. 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.Information Integration Solutions Center of Excellence Objective 2: The partition method must match the business requirements and stage functional requirements. 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. Remove Duplicates. All rights reserved. minimize repartitioning. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. No part of this publication may be reproduced. Note that in satisfying the requirements of this second objective. 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. 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. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. transcribed. Merge. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. but is not limited to: Aggregator. Objective 3: Unless partition distribution is highly skewed. and Sort stages. o Across jobs. Change Capture. or translated into any language in any form by any means without the written permission of IBM. This includes. It may also be necessary for Transformers and BuildOps that process groups of related records.

2Sorting and Hashing Advanced Example. To add aggregate columns to every detail row. a Transformer is used to extract data from a single header row of an input file. a new output column is defined on the header and detail links using a single constant value derivation. or translated into any language in any form by any means without the written permission of IBM. or through Auto partitioning): Parallel Framework Red Book: Data Flow Job Design July 17. 2006 70 of 179 © 2006 IBM Information Integration Solutions.4 Partitioning Examples In this section.Information Integration Solutions Center of Excellence 5. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator.4. 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. Additional. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly.2 Partitioning Example 2 – Use of Entire Partitioning In this example. more advanced partitioning and sorting examples are given in 12. the partitioning and sorting of this scenario can be optimized. transmitted. Because the Join and Aggregator use the same partition keys and sort order. 5. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. we’ll apply the partitioning methodology defined earlier to several example job flows. The output of the Aggregator is then sent to the second input of the Join. All rights reserved. and apply Same partitioning to the downstream links. on closer inspection. Using a “standard” solution.4.4. 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. Within the Transformer. 5. stored in a retrieval system. we can move the Hash partition and Sort before the Copy stage. transcribed. No part of this publication may be reproduced. .

so we will revisit this in the Sorting discussion. resulting in sequential processing. consider that the single header row is really a form of reference data. 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). or translated into any language in any form by any means without the written permission of IBM. For this reason. the link order in this example should be set so that the single header row is assigned to the Right input. it is not always the most efficient solution. stored in a retrieval system.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions . In order to process a large number of detail records. the single value join column will assign all rows to a single partition. 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. 2006 71 of 179 © 2006 IBM Information Integration Solutions. transcribed. Although functionally correct. Remembering that the degree of parallel operation is limited by the number of distinct values. the link order of the Inner Join is significant.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. the above solution has one serious limitation. . To optimize partitioning. there is no need to pre-sort the input to the Join. depending on the business requirements. transmitted. No part of this publication may be reproduced. All rights reserved. An optimized solution would be to alter the partitioning for the input links to the Join stage: .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 detail rows in the Left input will buffer to disk to prevent a deadlock.5. and may vary between job runs on the same Data Set. 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. All rights reserved. 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. there is a specialized example where the Round Robin collector may be appropriate. No part of this publication may be reproduced. 5. transmitted.Information Integration Solutions Center of Excellence If defined in reverse of this order. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5. (See Section 12. . 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. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition. For this reason.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially. or translated into any language in any form by any means without the written permission of IBM. there is one further detail in this example. Like partitioning methods. However. 5. the order of rows in an Auto collector is undefined. For advanced users.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.5. transcribed.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.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. stored in a retrieval system. Auto is the default collector method. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory.

5. 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. 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. The Sort Merge collector requires one or more key column(s) to be defined. Parallel Framework Red Book: Data Flow Job Design July 17. an Ordered collector will generate a sequential stream in sort order.5. in the same order.3 Ordered Collector An Ordered collector reads all rows from the first partition. 2006 73 of 179 © 2006 IBM Information Integration Solutions. 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.4 Sort Merge Collector If the input Data Set is sorted in parallel. All rights reserved. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter.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. use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. transmitted. No part of this publication may be reproduced. In this scenario. transcribed.6 Collecting Methodology Given the options for collecting data into a sequential stream. Row order is undefined for non-key columns. 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 used to sort the input Data Set in parallel. stored in a retrieval system. the Sort Merge collector will generate a sequential stream of rows in globally sorted order. as long as the Data Set has not been repartitioned or reduced.5. then reads all rows from the next partition until all rows in the Data Set have been collected. through aggregation). 5. . 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. Ordered collectors are generally only useful if the input Data Set has been Sorted and Range partitioned on the same key column(s).

a sort is required after repartitioning. Other stages (for example. ordered result set. 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.1 Partition and Sort Keys Using the parallel Sort within DataStage Enterprise Edition: . To restore row order and groupings. transmitted. when data is re-partitioned. the process of sorting data uses one primary key column and.Information Integration Solutions Center of Excellence 6 Sorting Traditionally. sorting on primary key LName (ascending). All rights reserved. Instead. optionally. Change Apply. as illustrated in the following example. 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. 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. No part of this publication may be reproduced. 2006 74 of 179 © 2006 IBM Information Integration Solutions. Change Capture. assigning rows with the same key column values to the same partition . The order of key columns determines the sequence and groupings in the result set. sorting is most often needed to establish order within specified groups of data. . the previous input Data Set is partitioned on LName and FName columns. In the following example. sort order is not maintained. based on one or more key column(s) NOTE: By definition. transcribed. Given a 4-node configuration file. Join. Merge) require pre-sorted groups of related records. stored in a retrieval system. This is the method the SQL databases use for an ORDER BY clause. in most cases there is no need to globally sort data to produce a single sequence of rows. or translated into any language in any form by any means without the written permission of IBM. Sort Aggregator. For example.Partitioning: is used to gather related records. 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. 6. one or more secondary key column(s) to generate a sequential. This sort can be done in parallel.Sorting: is used to establish group order within each partition.

2 Sorting and Hashing Advanced Example provides a more detailed discussion and example of partitioning and sorting. To satisfy these requirements we could:  Partition on CustID to group related records  Sort on OrderDate in Descending order  Remove Duplicates on CustID. 6. in general it is best to use a two step process: . or translated into any language in any form by any means without the written permission of IBM. Using Remove Duplicates.2 Complete (Total) Sort If a single. with Duplicate To Retain=First 12. Let’s say that an input Data Set consists of order history based on CustID and Order Date. 2006 75 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. transmitted. stored in a retrieval system. sequential ordered result is needed. 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). we want to select the most recent order for a given customer. 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. transcribed. .4.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. No part of this publication may be reproduced. For example.partition and parallel Sort on key column(s) Parallel Framework Red Book: Data Flow Job Design July 17.

All rights reserved.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). 2006 76 of 179 © 2006 IBM Information Integration Solutions. transcribed. links that have sort defined will have a Sort icon in addition to the partitioning icon. .3.1 Link Sort Sorting on a link is specified on the Input/Partitioning stage options. although the Enterprise Edition engine may insert a sort if required). or both . Within the Designer canvas.Information Integration Solutions Center of Excellence - use a Sort Merge collector on these same key column(s) to generate a sequential. When specifying key column(s) for partitioning. The stand-alone sort offers more options. (Sorting on a link is not available with Auto partitioning. both methods use the same internal sort package (the tsort operator). transmitted. partitioning. when specifying a keyed partitioning method.sort direction: Ascending or Descending . but is easier to maintain in a DataStage job. 6. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. By default. but as a separate stage makes job maintenance slightly more complicated. as there are fewer stages on the design canvas. No part of this publication may be reproduced.Sorting character set: ASCII (default) or EBCDIC (strings) Parallel Framework Red Book: Data Flow Job Design July 17. The Link sort offers fewer options. 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: . the standalone Sort stage is used to specify the Sort Key mode for partial sorts.key column usage: sorting.case sensitivity (strings) . ordered result set This is similar to the way parallel database engines perform their parallel sort operations. 6. Most often. the “Perform Sort” option is checked. use the Link sort unless a specific option is needed on the stand-alone Stage. In general.

This requires some additional overhead in the sort algorithm. the Sort Key Mode is most frequently used. and can often be performed in memory (depending on the size of the new subsort groups). It is used when resorting a sub-grouping of a previously sorted input Data Set. Always specify “DataStage” Sort Utility. which is significantly faster than a “UNIX” sort. instead of performing a complete Sort. For this reason. This “subsort” uses significantly less disk space and CPU resource. disable Stable sort unless needed. 2006 77 of 179 © 2006 IBM Information Integration Solutions.3. All rights reserved. - 6.5 Sub-Sorts Within the standalone Sort stage. or translated into any language in any form by any means without the written permission of IBM. transcribed. stored in a retrieval system. .4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys.Information Integration Solutions Center of Excellence - Position of nulls in the result set (for nullable columns) Figure 36: Specifying Link Sort options 6. transmitted. NOTE: The Sort Utility option is an artifact of previous releases. 6. 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. 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. Parallel Framework Red Book: Data Flow Job Design July 17.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. Specifically. the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization. No part of this publication may be reproduced.

as shown in the following example: Figure 38: Sort Key Mode property To successfully perform a subsort. If the input data does not match the key column definition for a subsort.p0] node2[op2. while new sort keys are specified with the “Sort” key mode property. transmitted. DataStage Enterprise Edition inserts sort operators as necessary to ensure correct results.2 Understanding the Parallel Job Score) can be used to identify automatically-inserted sorts. 2006 78 of 179 © 2006 IBM Information Integration Solutions. . 6. key={value=FirstName}}(0)) on nodes ( node1[op2. The parallel job score (see 12.Information Integration Solutions Center of Excellence To resort based on a sub-grouping.6 Automatically-Inserted Sorts By default. And. stored in a retrieval system. Re-used sort keys are specified with the “Don’t Sort (Previously Sorted)” property. Parallel Framework Red Book: Data Flow Job Design July 17. Enterprise Edition inserts sorts before any stage that requires matched key values or ordered groupings of (Join. No part of this publication may be reproduced. the job will abort.p1] node3[op2. Merge. Sorts are only inserted automatically when the flow developer has not explicitly defined an input sort. as shown in this score fragment: op1[4p] {(parallel inserted tsort operator {key={value=LastName}.p2] node4[op2.p3] )} Typically. the key column order for these keys must match the key columns and order defined in the previously-sorted input Data Set. All rights reserved. without gaps between them. keys with “Don’t Sort (Previously Sorted)” property must be at the top of the list. Remove Duplicates.4. transcribed. Sort Aggregator). all key columns must still be defined in the Sort stage. or translated into any language in any form by any means without the written permission of IBM.

If the sort operation can be performed in memory (as is often the case with a subsort) then no disk I/O is performed. memory. All rights reserved. stored in a retrieval system. transmitted. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be Parallel Framework Red Book: Data Flow Job Design July 17.4: Partitioning Examples. There are two ways to prevent Enterprise Edition from inserting an un-necessary sort: a) Insert an upstream Sort stage on each link. each sort uses 20MB of memory per partition for its memory buffer. To perform a sort. 2006 79 of 179 © 2006 IBM Information Integration Solutions. and disk resources. 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. transcribed. By default. This value can be changed for each standalone Sort stage using the “Restrict Memory Usage” option (the minimum is 1MB/partition). Revisiting the partitioning examples in Section 5. No part of this publication may be reproduced. Create Key Change Column. On a global basis.7 Sort Methodology Using the rules and behavior outlined in the previous section. rows in the input Data Set are read into a memory buffer on each partition. aborting the job if data is not in the required sort order. . 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. inserted sorts can be a significant performance impact if they are not necessary. 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. Create Cluster Key Change Column. sequential ordered result set use a parallel Sort and a Sort Merge collector 6. This will verify sort order but not actually perform a sort. 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.Information Integration Solutions Center of Excellence While ensuring correct results. or translated into any language in any form by any means without the written permission of IBM.8 Tuning Sort Sort is a particularly expensive task within DataStage Enterprise Edition which requires CPU.

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

No part of this publication may be reproduced. performance penalty of conversion. Need to share data with an external Parallel SAS application. stored in a retrieval system. Cannot write in parallel. transcribed. Data Sets store data in partitioned form. are summarized below: File Stage Recommended Usage Read and write standard files in a single format.2 Data Set Usage Parallel Data Sets are the persistent (on-disk) representation of the in-memory data structures of DS/EE. 2006 81 of 179 © 2006 IBM Information Integration Solutions. (Requires SAS connectivity license for DataStage. and any limitations. transmitted. Data Sets can only be read from and written to using a DataStage parallel job. can only be read from / written to by DS/EE or Parallel SAS. Requires Parallel SAS. Complex Flat File Data Set File Set SAS Parallel Data Set Lookup File Set No DS/EE file stage supports “update” of existing records. Some stages (parallel Data Set) support “Append” to add new records to an existing file. If data is Parallel Framework Red Book: Data Flow Job Design July 17. Slightly higher overhead than Data Set. Can only be used as reference link on a Lookup stage. 7.1 Which File Stage to Use DataStage/EE offers various stages for reading from and writing to files. As such. Intermediate storage between DataStage parallel jobs. . Can only be written – contents cannot be read or verified.) Rare instances where lookup reference data is required by multiple jobs and is not updated frequently. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. Data Sets provide maximum performance for reading and writing data from disk. Limitations Cannot write to a single file in parallel. Need to share information with external applications. performance penalty of format conversion. using the internal format of the parallel engine. Can only be read from and written to by DataStage parallel jobs or orchadmin command. Recommendations for when to use a particular stage. such as mainframe sources with COBOL copybook file definitions. but this is not recommended as it imposes risks for failure recovery. can write in parallel (generates multiple segment files). In general. 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. does not support hierarchical data files. However.Information Integration Solutions Center of Excellence 7 File Stage Usage 7.

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

or by using a File Pattern). the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. It is important to use ROUNDROBIN partitioning (or other partitioning appropriate to downstream components) to evenly distribute the data in the flow. the Sequential File (export operator) stage buffers its writes to optimize performance. • 7. Note that this document is required. transcribed.3. No part of this publication may be reproduced.3. 2006 83 of 179 © 2006 IBM Information Integration Solutions.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. but there is a small performance penalty associated with increased I/O. each file’s data is read into a separate partition.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. or translated into any language in any form by any means without the written permission of IBM. instead of statically through Table Definitions. as shown below: Parallel Framework Red Book: Data Flow Job Design July 17. making the entire downstream flow run sequentially (unless it is later repartitioned). . transmitted. When multiple files are read by a single Sequential File stage (using multiple files.5 Sequential File (Export) Buffering By default. Using the Schema File option allows the format of the source file to be specified at runtime.3. stored in a retrieval system. The format of the Schema File. 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. including Sequential File import / export format properties is documented in the Orchestrate Record Schema manual. 7. 7.4 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.7 Reading and Writing Nullable Columns When reading from or writing to Sequential Files or File Sets.3. the buffers are always flushed to disk. It is also important to remember that this setting will apply to all Sequential File stages in the data flow. SAME will read the entire file into a single partition. When a job completes successfully. All rights reserved. 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. Setting this value to a low number (such as 1) is useful for realtime applications. 7.

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

or OCCURS clauses. A complex flat file may contain one or more GROUPs. All rights reserved.10 Tuning Sequential File Performance On heavily-loaded file servers or some RAID/SAN array configurations. in some disk array configurations. . these extra characters will be silently truncated. NOTE: The Complex Flat File stage cannot read from sources with OCCURS DEPENDING ON clauses. 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. (This is an error in the DataStage documentation. 7. It does not write to MVS datasets. the stage allows you to write data to one or more complex flat files. 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.) When used as a target. REDEFINES. These settings specify the size of the read (import) and write (export) buffer size in Kbytes. 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. but you can configure the stage to execute sequentially if it is only reading one file with a single reader. Increasing this size may improve performance. transcribed. transmitted. When used as a source.Information Integration Solutions Center of Excellence Varchar length. with a default of 128 (128K).3. the stage allows you to read data from one or more complex flat files. Finally. stored in a retrieval system. Complex Flat File source stages execute in parallel mode when they are used to read multiple files. the environment variables $APT_IMPORT_BUFFER_SIZE and $APT_EXPORT_BUFFER_SIZE can be used to improve I/O performance.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. 7. including MVS datasets with QSAM and VSAM files. 2006 85 of 179 © 2006 IBM Information Integration Solutions.

leading separate. native binary binary. transcribed. 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. trailing zoned.y] decimal[x+y.y] decimal[x+y. trailing separate.y] decimal[x+y. leading Parallel Framework Red Book: Data Flow Job Design July 17. transmitted.y] decimal[x+y.4. native binary binary. trailing zoned.1 CFF Stage Data Type Mapping When you work with mainframe data using the CFF stage.y] or string[x+y] decimal[x+y. stored in a retrieval system.y] decimal[x+y.y] or string[x+y] decimal[x+y. or translated into any language in any form by any means without the written permission of IBM.y] sfloat dfloat ustring[n] ustring[max=n] subrec Internal Options packed packed zoned 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. All rights reserved. native binary binary. . 2006 86 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 7. native binary binary. 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.

1.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints.1. entries are placed in the Director job log. When rows are rejected by a Transformer. and to use other stages (such as Copy) when derivations are not needed. 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. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. if you intend to use a nullable column within a Transformer derivation or output link constraint. Because the DS/EE Transformer runs in parallel. Even if the target column in an output derivation allows nullable results. it is important to minimize the number of transformers. All rights reserved.8. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17. 2006 87 of 179 © 2006 IBM Information Integration Solutions. For this reason. 8. undefined. 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. To create a Transformer reject link in Designer.1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. 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.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer.4: Parallel Transformer stages for guidelines on Transformer stage usage.col) Then “” Else link. transmitted. For example.Information Integration Solutions Center of Excellence 8 Transformation Languages 8.col Note that if an incoming column is only used in an output column mapping. For this reason. See Section 3. transcribed. Always include reject links in a parallel Transformer. by definition. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link. 8. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. No part of this publication may be reproduced. This makes it easy to identify reject conditions (by row counts). .

6 -> -1 floor Rounds towards negative infinity. internal decimal results are rounded to the nearest applicable value. or performing parallel derivations. transcribed. there are times when internal decimal variables need to be generated in order to perform the evaluation. the PadString function uses the length of the source type. For example. Examples: 1.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment.1.4 -> -2 round_inf Rounds or truncates towards nearest representable value.5-> 2. For example. By default.1. 2006 88 of 179 © 2006 IBM Information Integration Solutions. transmitted. By default. the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range. the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. Within the link constraints dialog box. it is possible that valid rows will not have been flushed from Sequential File (export) buffers. -1. not the target. If more precision is required. Thus. up to a maximum precision of 255 and scale of 125. -1. these internal decimal variables will have a precision and scale of 38 and 10.3. -1. the incoming column must be type VarChar before it is evaluated in the Transformer.5: Sequential File (Export) Buffering).4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule. Create a new output link that will handle rows that match the abort rule. . Therefore. All rights reserved. -1. 1. 8. breaking ties by rounding positive values toward positive infinity and negative values toward negative infinity. or translated into any language in any form by any means without the written permission of IBM. 1).4 -> 1.4 -> -1.4 -> 2. TrimLeadingTrailing(string) works only if string is a VarChar field. When generating a sequence of numbers in parallel. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7. stored in a retrieval system. apply the abort rule to this output link. 8. and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example. or committed to database tables. it is important to make sure the type conversion is done before a row reaches the Transformer. 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.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition.1.6 ->1. No part of this publication may be reproduced. 8. Since the Transformer will abort the entire job flow immediately. Examples: 1. Examples: 1.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17.

the output links are also evaluated in the order in which they are displayed. stored in a retrieval system. it can be seen that there are certain constructs that would be inefficient to include in output column derivations. transmitted. In this case.col[1. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. Examples: 1. From this sequence. the evaluation of the substring of DSLINK1. 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. All rights reserved. -1. 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.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations.3] = “001”) THEN . or translated into any language in any form by any means without the written permission of IBM. 8. round or truncate to the scale size. By doing this.56 ->-1..3] is evaluated for each column that uses it. Similarly. The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value. No part of this publication may be reproduced. This can be made more efficient by moving the substring calculation into a stage variable.5. In this case. as they would be evaluated once for every output column that uses them.56 -> 1. it is useful to understand what items get evaluated and when. transcribed.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign.col[1. 2006 89 of 179 © 2006 IBM Information Integration Solutions. the substring is evaluated just once for every input row. suppose multiple columns in output links want to use the same substring of an input column. .1. For example. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example.5. then the following test may appear in a number of output columns derivations: IF (DSLINK1.

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

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

direction defaults to end. and timestamp fields. 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. The character argument is the character to remove. 8. time. transmitted. By default. .2. Justify has no affect when the target string has variable length. . transcribed.value) where: . this is NULL.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. the syntax is: destField[:dataType] = make_null(sourceField. To convert from an out-of-band null to an in-band null (value) representation within Modify. and justify defaults to begin. direction. The value of the direction and justify arguments can be either begin or end.value is the value you wish to represent a null in the output. . 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. The destField is converted from an Orchestrate out-of-band null to a value of the field’s data type.value is the value of the source field when it is null.destField is the destination field’s name. for decimal.dataType is its optional data type.Information Integration Solutions Center of Excellence 8. To convert from an in-band null to an out-of-band null. All rights reserved. the syntax is: destField[:dataType] = handle_null (sourceField.sourceField is the source field’s name . with the following syntax: stringField=string_trim[character. .destField is the destination field’s name. these characters are retained when the fixed-length string is then converted back to a variable-length string. value can be a string. NOTE: The DataStage Parallel Job Developers Guide gives incorrect syntax for converting an out-of-band null to an in-band null (value) representation. By default. . stored in a retrieval system.2 Modify and String Trim The function string_trim has been added to Modify. .2. For a numeric field value can be a numeric value. date. use it if you are also converting types. use it if you are also converting types. 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.value) where: . string.dataType is its optional data type. 2006 92 of 179 © 2006 IBM Information Integration Solutions.sourceField is the source field’s name.

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

This is also handy with Lookups that have multiple reference links. and to retest if the relative volumes of data change dramatically. Right Outer. For example. when a match does not occur. Join). Join vs. . Each lookup reference requires a contiguous block of shared memory. an OUTER join operation must be performed. the default value for a Varchar is an empty string (“”). No part of this publication may be reproduced. Oracle Enterprise.1. 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. 9. This is most easily done by inserting a Copy stage and mapping a column from NON-NULLABLE to NULLABLE. If un-matched rows must be captured or logged. In most cases. and ODBC Enterprise stages) to scenarios where the number of input rows is significantly smaller (for example. the default value for an Integer is zero.7: Database Sparse Lookup vs. with a constant value. transmitted. It is best to test both the Sparse and Normal to see which actually performs best. 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. 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. During an Outer Join. or both links in the case of Full Outer) are output regardless of match on key values. transcribed. stored in a retrieval system. If the Data Sets are larger than available memory resources.1 Lookup vs. (see Section 10. For this reason. 2006 94 of 179 © 2006 IBM Information Integration Solutions. This isolates your match/no-match logic from any changes in the metadata. all rows on an outer link (for example. Limit the use of database Sparse Lookups (available in the DB2 Enterprise. Left Outer. to each of the inner links and test that column for the constant after you have performed the join.Information Integration Solutions Center of Excellence 9 Combining Data 9. In an OUTER join scenario.2 Capturing Unmatched Records from a Join The Join stage does not provide reject handling for unmatched records (such as in an InnerJoin scenario). A Transformer stage can be used to test for NULL values in unmatched columns. All rights reserved. 1:100 or more) than the number of reference rows. Sparse Lookups may also be appropriate for exception-based processing when the number of exceptions is a small fraction of the main input data. and the default value for a Char is a string of padchar characters equal to the length of the Char column. it is best to use a Column Generator to add an ‘indicator’ column. then Enterprise Edition inserts “default” values based on the data type. care must be taken to change the column properties to allow NULL values before the Join. the JOIN or MERGE stage should be used.

which maintains the results of each keycolumn value/aggregation pair in memory. transcribed. the Hash Aggregator should only be used when the number of distinct key values is small and finite.3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings. Parallel Framework Red Book: Data Flow Job Design July 17.2 Aggregation Data Type By default. To perform a total aggregation. 9. Because each key value/aggregation requires approximately 2K of memory. Unlike the Hash Aggregator.3.3 The Aggregator Stage 9.Information Integration Solutions Center of Excellence 9. 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. the default Aggregation Method is set to Hash. 2006 95 of 179 © 2006 IBM Information Integration Solutions. An exception to this is financial calculations which should be done in decimal to preserve appropriate precision.3. stored in a retrieval system. but only maintains the calculations for the current group in memory. 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. 9. transmitted.3. the output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double). set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. 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. To aggregate in decimal precision. the Sort Aggregator requires presorted data. All rights reserved. . 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. The Sort Aggregation Method should be used when the number of key values is unknown or very large.1 Aggregation Method By default. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property.

All rights reserved. and Teradata). Oracle. or translated into any language in any form by any means without the written permission of IBM.1 Database stage types DataStage Enterprise Edition offers database connectivity through native parallel and plug-in stage types. For some databases (DB2. 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. . 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 17.1 Database development overview This section is intended to provide guidelines appropriate to accessing any database within DataStage Enterprise Edition. stored in a retrieval system. transmitted. Informix. 2006 96 of 179 © 2006 IBM Information Integration Solutions. 10. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 10 Database Stage Guidelines 10. Subsequent sections provide database-specific tips and guidelines. Teradata API) are visible in the default DataStage Designer palette.1. transcribed. You may need to customize the palette to add hidden stages.

the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server. and cannot span multiple servers in a clustered or Grid configuration. operating system. Unlike the database-specific parallel stages. Because of their tight integration with database technologies. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect. 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). Native parallel stages always pre-query the database for actual runtime metadata (column names.1 Native Parallel database stages In general. Furthermore. and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied. 10. not position in the stage Table Definitions. 10. and version. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform. attributes). No part of this publication may be reproduced. . for maximum parallel performance.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. Parallel Framework Red Book: Data Flow Job Design July 17. Because plug-in stage types cannot read in parallel. they should only be used when it is not possible to use a native parallel stage. native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform. plug-in database stages match columns by order. transcribed.1. For example. All rights reserved. the ODBC Enterprise stage cannot interface with database-specific parallel load technologies. transmitted. types. or translated into any language in any form by any means without the written permission of IBM. This allows Enterprise Edition to match return columns by name. 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. not name.2 ODBC Enterprise stage In general. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. the native parallel stages often have more stringent connectivity requirements than plug-in stages. From a design perspective.Information Integration Solutions Center of Excellence 10.1. care must be taken to assign the correct data types in the job design.1.1. scalability. 2006 97 of 179 © 2006 IBM Information Integration Solutions.1. stored in a retrieval system.1. the DataDirect ODBC Driver Manager is also included. so Table Definitions must match the order of columns in a query. Because there are exceptions to this rule (especially with Teradata). specific guidelines of when to use various stage types are provided in the database-specific topics in this section. However. On UNIX.

transcribed. 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.1. Informix Enterprise or Oracle Enterprise stages. For each native parallel database stage: . 2006 98 of 179 © 2006 IBM Information Integration Solutions. transmitted. No part of this publication may be reproduced. data types.2 Database Metadata 10.Information Integration Solutions Center of Excellence 10. 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. or translated into any language in any form by any means without the written permission of IBM. 10.2. stored in a retrieval system. .columns of the database row correspond to columns of a DS/EE record .rows of the database result set correspond to records of a DS/EE Data Set . All rights reserved. This allows the database stages to match return values by column name instead of position. use orchdbutil to import metadata to avoid type conversion issues.both DS/EE and relational databases support null values.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 .1. However. and selecting “Import from Database Table” option in the wizard as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17.1 Runtime metadata 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.1. nullability) and partitioning scheme (in some cases) of the source or target table. the DS/EE native parallel database stages always “pre-query” the database source or target to determine the actual metadata (column names.2.2 Metadata Import When using the native parallel DB2 Enterprise. Database-specific data type mapping tables are included in the following sections.

For example. No part of this publication may be reproduced. 10. 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. transcribed.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. The alias name(s) should then be added to the Table Definition within DataStage.1.3 Defining Metadata for Database Functions When using database functions within a SQL SELECT list in a Read or Lookup. stored in a retrieval system. the DataStage Designer will automatically populate the select list based on the stage’s output column definition. there may be cases where user-defined functions or logic need to be executed on the database server. it will be easier to use the corresponding orchdbutil command-line utility from the DataStage server machine. 2006 99 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.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. However. If the connection is successful. the following SQL assigns the alias Total to the calculated column: SELECT store_name. .1. instead of using an unqualified “Table” or SQL “SELECT *” read. you will see a window with the result columns and data.2. orchdbutil can be scripted to automate the process of importing a large number of tables. As a command. similar to the illustration on the right: Parallel Framework Red Book: Data Flow Job Design July 17. When importing a large number of tables. Test the connection using View Data button. it is best to explicitly specify column names on all source database stages. always specify the “Select List” subproperty. For “Table” read method. 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.1. 10. transmitted. 10. For “Auto-Generated” SQL. it is important to use SQL aliases to explicitly name the calculated columns so that they can be referenced within the DataStage job.3 Optimizing Select Lists For best performance and optimal memory usage. or translated into any language in any form by any means without the written permission of IBM.

6 Database OPEN and CLOSE Commands The native parallel database stages provide options for specifying OPEN and CLOSE commands. transcribed.Information Integration Solutions Center of Excellence Figure 43: Sample View Data Output If the connection fails. OPEN and CLOSE are not offered by plugin database stages. a Lookup stage (or Join stage. depending on data volume) can be used to identify existing rows before they are inserted into the target table. it is important to separate the transformation process from the database write (Load or Upsert) operation. 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. including databasespecific options (tablespace.1. 2006 100 of 179 © 2006 IBM Information Integration Solutions. For example. and doing so may violate data-management (DBA) policies. logging. it is not a good idea to let DataStage generate target tables unless they are used for temporary storage. and you will be prompted to view additional detail. In general. the OPEN command can be used to create a target table. constraints. transmitted. 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. All rights reserved. 10. Figure 44: View Additional Error Detail 10. an error message may appear. stored in a retrieval system. and the CLOSE command could be used to select all rows from the temporary table and insert into a final target table. There are limited capabilities to specify Create table options in the stage. Parallel Framework Red Book: Data Flow Job Design July 17. etc) not possible with the “Create” option. the OPEN command could be used to create a temporary table. Subsequent job(s) should read this Data Set and populate the target table using the appropriate database stage and write method. As a further optimization.5 Designing for Restart To enable restart of high-volume jobs. As another example. .1. the results should be landed to a parallel Data Set. After transformation. 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.

In most cases. Parallel Framework Red Book: Data Flow Job Design July 17. it is faster to use a DataStage JOIN stage between the input and DB2 reference data than it is to perform a “Sparse” Lookup.1. transcribed. When specifying a user-defined OPEN command. ODBC Enterprise. Join Data read by any database stage can serve as the reference input to a Lookup operation. 10.8 Appropriate Use of SQL and DataStage When using relational database sources. This minimizes impact on network and memory resources. and ongoing application maintenance costs. By default.1. While there are extreme scenarios when the appropriate technology choice is clearly understood. this reference data is loaded into memory like any other reference link (“Normal” Lookup).7 Database Sparse Lookup vs. a default OPEN statement places a shared lock on the source. when reading from DB2. When directly connected as the reference link to a Lookup stage. use a SQL filter (WHERE clause) to limit the number of rows sent to the DataStage job. Further details are outlined in the respective database sections of the Orchestrate Operators Reference which is part of the Orchestrate OEM documentation.Information Integration Solutions Center of Excellence It is important to understand the implications of specifying a user-defined OPEN and CLOSE command. or translated into any language in any form by any means without the written permission of IBM. a Sparse Lookup may be appropriate. there is often a functional overlap between SQL and DataStage functionality. and Oracle Enterprise stages allow the lookup type to be changed to “Sparse”. the optimal implementation involves leveraging the strengths of each technology to provide maximum throughput and developer productivity. The following guidelines can assist with the appropriate use of SQL and DataStage technologies in a given job flow: • When possible. No part of this publication may be reproduced. sending individual SQL statements to the reference database for each incoming Lookup row. 2006 101 of 179 © 2006 IBM Information Integration Solutions. and leverages the database capabilities. Although it is possible to use either SQL or DataStage to solve a given business problem. Sparse Lookup is only available when the database stage is directly connected to the reference link. IMPORTANT: The individual SQL statements required by a “Sparse” Lookup are an expensive operation from a performance perspective. For scenarios where the number of input rows is significantly smaller (for example. transmitted. this lock is not sent – and should be specified explicitly if appropriate. there may be “gray areas” where the decision should be made based on factors such as developer productivity. metadata capture and re-use. 10. . stored in a retrieval system. the DB2/UDB Enterprise. 1:100 or more) than the number of reference rows in a DB2 or Oracle table. All rights reserved. For example. with no intermediate stages.

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

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.2 DB2 Guidelines 10. the DB2 hardware/UNIX/software platform must match the hardware/software platform of the DataStage ETL server.1. transmitted. consult the DataStage documentation (DataStage Parallel Job Developers Guide. and DataStage Plug-In guides) 10. upsert. 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. 3 A patched version of the ODBC Enterprise stage allowing parallel read is available from IBM IIS Support for some platforms. This will only work when the authentication mode of the database on the remote instance is set to “client authentication”. lookup.Information Integration Solutions Center of Excellence 10. For this reason. 2006 103 of 179 © 2006 IBM Information Integration Solutions. 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.) Furthermore. transcribed.1 DB2/UDB Enterprise stage Enterprise Edition provides native parallel read. If you use the stage in this way. Check with IBM IIS Support for availability. 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.2. 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. the “client authentication” configuration of a remote instance is not recommended. this was also called “DB2 EEE”. (Before DB2 v8. stored in a retrieval system. and load capabilities to parallel DB2 databases on UNIX using the native parallel DB2/UDB Enterprise stage. No part of this publication may be reproduced.1 DB2 Stage Types DataStage Enterprise Edition provides access to DB2 databases using one of 5 stages.2. . All rights reserved. The DB2/UDB Enterprise stage requires DB2 Enterprise Server Edition on UNIX with Data Partitioning Facility (DPF) option.

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). 2006 104 of 179 © 2006 IBM Information Integration Solutions.1. databases on UNIX platforms that differ from the platform of the DataStage ETL server. parallel component the DB2/UDB Enterprise stage is designed for maximum performance and scalability. Sparse Lookup is not supported through the DB2/API stage. 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.2 Connecting to DB2 with the DB2/UDB Enterprise Stage Create a Parallel job and add a DB2/UDB Enterprise stage. using the same data partitioning as the referenced DB2 tables. stored in a retrieval system. or DB2 databases on Windows or Mainframe platforms (except for the “Load” stage against a mainframe DB2 instance which is not supported). Add the following properties: Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. the ability to write in parallel may be limited by the table and index configuration set by the D2 database administrator.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 v8.2. 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. These goals are achieved through tight integration with the DB2 RDBMS. when used as data sources. These stages also provide connectivity to non-UNIX DB2 databases. By facilitating flexible connectivity to multiple types of remote DB2 database servers. including direct communication with each DB2 database node. No part of this publication may be reproduced. Because each plug-in invocation will open a separate connection to the same target DB2 database table. “DB2 EE”). 10.Information Integration Solutions Center of Excellence As a native. All rights reserved.writing to DB2 in parallel (where appropriate). it may be possible to write to a DB2 target in parallel. 10. .2. Furthermore. the use of DataStage plug-in stages will limit overall performance and scalability. plug-in stages cannot read from DB2 in parallel. transmitted. and reading from .

3 Configuring DB2 Multiple Instances in One DataStage Job Although it is not officially supported. 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. to identify the instance name of the DB2 server. • Client Alias DB Name. 2006 105 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Otherwise use the DB2 environment variable.Information Integration Solutions Center of Excellence Figure 45: DB2/UDB Enterprise stage properties For connection to a remote DB2/UDB instance.2. All rights reserved. Enter the user name for connecting to DB2. transmitted. 10. DataStage assumes you require remote connection. DB2INSTANCE. • Server. or translated into any language in any form by any means without the written permission of IBM. .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. If you set this property. 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. Once this variable is set. Optionally set this to the remote server database name. 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. 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.cfg file can July 17. • Password. it is possible to connect to more than one DB2 instance within a single job. Set this to the DB2 client’s alias database name for the remote DB2 server database. Enter the password for connecting to DB2. Single stream . Set this to the DB2 client instance name. you will need to set the following properties on the DB2/UDB Enterprise stage in your parallel job: • Client Instance Name. 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. 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. Optionally set this to the instance name of the DB2 server. it will try to use it for each of the connections in the job. • User. • Database. Otherwise use the environment variables $APT_DBNAME or $APT_DB2DBDFT to identify the database. transcribed. stored in a retrieval system. No part of this publication may be reproduced. This is required only if the client’s alias is different from the actual name of the remote server database. Since a db2nodes.

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

. 10.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. DECIMAL[p. No part of this publication may be reproduced. stored in a retrieval system.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). The original names are used in generated SQL statements. and you should use them if entering SQL in the job yourself.2. DB2 Data Type CHAR(n) CHARACTER VARYING(n. The DB2/UDB Enterprise stage converts DB2 data types to Enterprise Edition data types. Alternatively. Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. avoid hand editing (this minimizes the risk of mistakes or confusion). the result is a timestamp field.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. transmitted.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. examine the DDL for each schema to be accessed. 2006 107 of 179 © 2006 IBM Information Integration Solutions.r) NVARCHAR(n. or translated into any language in any form by any means without the written permission of IBM. the result is a time field.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.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n. All rights reserved. decimal[p. Once the table definition is loaded. as shown in the following table. however. 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. transcribed. They are also used in derivations and expressions. the internal column names are displayed rather than the original DB2 names both in table definitions and in the Data Browser.

That is. In these instances. DB2 Enterprise stage should not include the Client Instance Name property. the DB2 Load DataStage job must be re.run in Truncate mode to clear the load pending state.2. communicating directly with each database node. In this configuration. and the availability of new rows to other applications. and writing to a DB2 Enterprise Server Edition database with Database Partitioning Feature (DPF) • As a native. transmitted. and the target table(s) may be accessed by other users. the overhead of routing information through a remote DB2 coordinator may be significant. The DB2 load operator performs a non-recoverable load. stored in a retrieval system. . 10.7 DB2/UDB Enterprise stage options The DB2/UDB Enterprise (native parallel) stage should be used for reading from. the DB2/UDB Enterprise stage is designed for maximum performance and scalability against very large partitioned DB2 UNIX databases. the DB2/UDB Enterprise stage offers the choice of SQL (insert / update / upsert / delete) or fast DB2 loader methods. database log usage. Parallel Framework Red Book: Data Flow Job Design July 17.2. No part of this publication may be reproduced. All operations are logged to the DB2 database log. a) The Write Method (and corresponding insert / update / upsert / delete) communicates directly with the DB2 database nodes to execute instructions in parallel. When writing to a DB2 database in parallel. if the load operation is terminated before it is completed. • DB2/UDB Enterprise stage is tightly integrated with the DB2 RDBMS. 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). reading from and writing to DB2 in parallel (where appropriate). forcing the DB2 Enterprise stages on each ETL server to communicate directly with their local DB2 coordinator. The choice between these methods depends on required performance. the contents of the table are unusable and the tablespace is left in a load pending state. performing lookups against. and recoverability. when using user-defined SQL without partitioning against large volumes of DB2 data. During the load operation. All activity in the z/OS environment always goes through the DB2 coordinator node so parallelism differs slightly depending on how DB2 is accessed.8 Performance Notes In some cases. 10. and using the same data partitioning as the referenced DB2 tables.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. b) The DB2 Load method requires that the DataStage user running the job have DBADM privilege on the target DB2 database. 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. transcribed.Information Integration Solutions Center of Excellence 10. 2006 108 of 179 © 2006 IBM Information Integration Solutions. Time and row-based commit intervals determine the transaction size.2. In this scenario. or translated into any language in any form by any means without the written permission of IBM. parallel component. All rights reserved.

partitioning index name(s). Table T is in tablespace TS and TS is partitioned into 3 partitions on Col1 (limits: F. This information determines the number of db2read operators that the conductor builds into the score and the queries that they execute. . These values are in turn used to read the SYSTABLEPART table to retrieve the number of partitions. transcribed. 40).Information Integration Solutions Center of Excellence When accessing a DB2 table using the Table read method. No part of this publication may be reproduced. T) and Col2 (10. This is illustrated in Figure 47. Parallel Framework Red Book: Data Flow Job Design July 17. P. and is controlled by the number of nodes in the configuration file. as illustrated in Figure 46: Figure 46: DB2 read on DataStage/USS For example. or translated into any language in any form by any means without the written permission of IBM. and the partition limit key value(s). the number of operators do not have to match to the number of partitions. 2006 109 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. 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. functions within the db2read operator are used to read the DB2 SYSTABLES table to retrieve the tablespace and database name for the table. transmitted. Finally the SYSKEYS and SYSCOLUMNS tables are read using the index name to get the associated column metadata (name and type). 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. 30.

transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. An example of an in-memory Normal Lookup is shown in Figure 48. transcribed. 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. 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. stored in a retrieval system.Information Integration Solutions Center of Excellence Figure 47: DB2 write on DataStage/USS On DataStage/USS Edition. No part of this publication may be reproduced. All rights reserved.

using the DB2 load utility in USS is different from non-z/OS environments. Figure 49: DB2 Sparse Lookup on DataStage/USS Finally. stored in a retrieval system. 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. 2006 111 of 179 © 2006 IBM Information Integration Solutions. we call a DB2 stored procedure called DSNUTILS. Since each of these queries must go through the DB2 coordinator node we can effectively ignore the level of parallelism specified for the table. or translated into any language in any form by any means without the written permission of IBM. we need to add a special resource statement in our configuration file to specify the MVS dataset name to use. Contrast the Normal Lookup with the way a Sparse Lookup is done as shown in Figure 49. transcribed. In order to invoke it from a DataStage/USS job. Since there is no sequential file stage associated with this MVS load file. where each lookup operator is issuing an SQL to DB2 for every row it processes. 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. transmitted. The LOAD utility has a second limitation in that data cannot be piped into it. All rights reserved. nor can it be read in from a USS HFS file.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. . Parallel Framework Red Book: Data Flow Job Design July 17. When the conductor creates the score. No part of this publication may be reproduced. 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. The DB2 LOAD utility is designed to run from JCL only.

All rights reserved.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. 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. stored in a retrieval system. transmitted. transcribed. 2006 112 of 179 © 2006 IBM Information Integration Solutions. .

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. No part of this publication may be reproduced.columns of the database row correspond to columns of a DS/EE record .Information Integration Solutions Center of Excellence 10.3 Informix Database Guidelines 10. the result is a timestamp field. stored in a retrieval system.r) NVARCHAR(n.s] where p is the precision and s is the scale The maximum precision is 32.2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date.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 . time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month. 2006 113 of 179 © 2006 IBM Information Integration Solutions.both DS/EE and Informix support null values. If the DATETIME starts with a year component. Informix Data Type CHAR(n) CHARACTER VARYING(n. The Informix Enterprise stage converts Informix data types to Enterprise Edition data types. All rights reserved.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: . the result is a date field. transmitted. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.rows of the database result set correspond to records of a DS/EE Data Set .s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. transcribed.3. decimal[p. DECIMAL[p.3. . the result is a time field. or translated into any language in any form by any means without the written permission of IBM. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. as shown in the following table.

columns of the database row correspond to columns of a DS/EE record .4. If you attempt to read a record larger than 32K. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.4 ODBC Enterprise Guidelines 10.both DS/EE and ODBC support null values. transmitted. It is important to verify the correct ODBC to Enterprise Edition data mapping. two underscore characters replace the unsupported character .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.Information Integration Solutions Center of Excellence 10.4. . 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.2 ODBC Enterprise stage Data Type Mapping ODBC data sources are not supported by the orcdbutil utility. Parallel Framework Red Book: Data Flow Job Design July 17. Enterprise Edition will return an error and abort your job. All rights reserved.names are translated exactly except when the external data source column name contains a character that DataStage does not support. No part of this publication may be reproduced. stored in a retrieval system.1 ODBC Enterprise Stage Column Names For each ODBC Enterprise stage: .s] decimal[p.rows of the database result set correspond to records of a DS/EE Data Set . In that case.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 . or translated into any language in any form by any means without the written permission of IBM. 2006 114 of 179 © 2006 IBM Information Integration Solutions.s] string[36] int32 int8 [0 or 1] decimal[p. transcribed.s] where p is the precision and s is the scale decimal[p.

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. 2006 115 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 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. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. transmitted. . All rights reserved.

stored in a retrieval system. All rights reserved. 2006 116 of 179 © 2006 IBM Information Integration Solutions. and the table configuration (row or page level lock mode if available). it may be possible to write to a target database in parallel using the ODBC Enterprise stage. a patch may be available through IBM IIS Support to support parallel reads through ODBC. Parallel reads through ODBC match the degree of parallelism in the $APT_CONFIG_FILE. the ODBC Enterprise stage does not support parallel read4. transcribed. No part of this publication may be reproduced. transmitted.4. . since this capability is not provided by the ODBC API. Depending on the target database. 4 On some platforms.3 Reading ODBC Sources Unlike other native parallel database stages. 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.Information Integration Solutions Center of Excellence 10.

names are translated exactly except when the Oracle source column name contains a character that DataStage does not support. and will generate an error at runtime 10.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p. Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p. No part of this publication may be reproduced. 2006 117 of 179 © 2006 IBM Information Integration Solutions. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage. or translated into any language in any form by any means without the written permission of IBM. the Oracle Enterprise stage reads sequentially from its source table or query. All rights reserved.2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. Enterprise Edition will return an error and abort your job.s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38. two underscore characters replace the unsupported character . stored in a retrieval system.columns of the database row correspond to columns of a DS/EE record .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.5.3 Reading from Oracle in Parallel By default.5.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 . If you attempt to read a record larger than 32K. Setting the partition table option to the specified table will enable parallel extracts from an Oracle source. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. .5 Oracle Database Guidelines 10. which are not heavily typed.Information Integration Solutions Center of Excellence 10. transmitted. transcribed.1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . This is particularly important for Oracle databases.rows of the database result set correspond to records of a DS/EE Data Set . The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition.5.both DS/EE and Oracle support null values. Examples include: .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. In that case.

the Oracle load will be done sequentially. stored in a retrieval system.5. 2006 118 of 179 © 2006 IBM Information Integration Solutions. transcribed. 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. 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. 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. PARALLEL=FALSE) also allows loading of indexed tables without index maintenance. Parallel Framework Red Book: Data Flow Job Design July 17. • Setting the environment variable $APT_ORACLE_LOAD_OPTIONS to “OPTIONS (DIRECT=TRUE. transmitted. When using this method. . rebuild). All rights reserved.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) 10.4 Oracle Load Options When writing to an Oracle table (using Write Method = Load). In this instance. 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.

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.

The SessionsPerPlayer option determines the number of connections each DataStage EE player opens to Teradata. RequestedSessions equals the maximum number of available sessions on the Teradata instance. where pa$$ is the example password: \’pa$$\’ 10. 10.6 Specifying Teradata Passwords with Special Characters Teradata permits passwords with special characters and symbols. To specify a Teradata password that contains special characters. the DB Options property specifies the connection string and connection properties in the form: user=username. this determines the number of DataStage players. . Teradata Enterprise stage uses 32K buffers.7. transmitted.7. No part of this publication may be reproduced.RequestedSessions=nn] where SesionsPerPlayer and RequestedSessions are optional connection parameters that are required when accessing large Teradata databases.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.7. the password must be surrounded by an “escaped” single quote as shown. Parallel Framework Red Book: Data Flow Job Design July 17. and/or RequestedSessions should be decreased. Setting the SessionsPerPlayer too low on a large system can result in so many players that the job fails due to insufficient resources. or translated into any language in any form by any means without the written permission of IBM. 2006 122 of 179 © 2006 IBM Information Integration Solutions.7 Teradata Enterprise Settings Within the Teradata Enterprise stage.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. transcribed. and will generate an error at runtime. By default. Aggregates and most arithmetic operators are not allowed in the SELECT clause of a Teradata Enterprise stage. 10. In that case SessionsPerPlayer should be increased. Indirectly. All rights reserved. By default. and hence the number of UNIX processes and overall system resource requirements of the DataStage job. 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. (Note that 64K buffers must be enabled at the Teradata server level). but this can be set to a value between 1 and the database vprocs.password=password[. stored in a retrieval system.SessionsPerPlayer=nn][.

transmitted.Information Integration Solutions Center of Excellence 10. That leaves 39 possible TDP names and is different than the convention used for non-channel attached systems. The first three characters must be TDP. On a network-attached system. the tdpid must be in the form TDPx. #. $. where x is 0-9. stored in a retrieval system. or @. 2006 123 of 179 © 2006 IBM Information Integration Solutions. To connect to a Teradata server. you must supply the client with the Teradata Director Program (TDP) identifier. 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. On MVS. also known as the tdpid. the tdpid is the host name of the Teradata server. All rights reserved. . No part of this publication may be reproduced. transcribed. A-Z (case insensitive).7.9 Teradata on USS On the USS platform the Teradata Enterprise Stage uses CLIv2 for channel-attached systems (OS/390 and z/OS).

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

tune. the log entry is sometimes preceded by a warning condition. it may not be possible to remove all warning messages generated by the EE engine. All rights reserved. No part of this publication may be reproduced. When a fatal error occurs. Use the Data Set Management tool (available in the Tools menu of DataStage Designer or DataStage Manager) to examine the schema.Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. it is possible for $OSH_PRINT_SCHEMAS to generate a log entry that is too large for DataStage Director to store or display. or translated into any language in any form by any means without the written permission of IBM.  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled). 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. July 17. This will place entries in the Director log with the actual runtime schema for every link using Enterprise Edition internal data types. stored in a retrieval system. transcribed. or promote a job from development into test or production. Set the environment variable $DS_PX_DEBUG to capture copies of the job score.    NOTE: For large jobs. 11. and internal Enterprise Edition log messages in a directory corresponding to the job name.5.  Enable the Job Monitoring Environment Variables detailed in Section 2. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. In some instances. To capture the full schema output in these cases. The default is 30 seconds. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. look at row counts. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. . These may indicate an underlying logic problem or unexpected data type conversion. But all warnings should be examined and understood. The general process for debugging a job is:  Check the Director job log for warnings. All fatal and warning messages should be addressed before attempting to debug. transmitted. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG .3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. generated osh.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide. and to manage source or target Parallel Data Sets.

It is useful to examine the generated OSH for debugging purposes. transcribed. 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. No part of this publication may be reproduced.4 Viewing the Generated OSH Within Designer. To enable viewing of generated OSH. or translated into any language in any form by any means without the written permission of IBM. and to understand internally what is running. stored in a retrieval system. It is also not useful for files of non-delimited fixed-length record format. so if the file has any binary columns. NOTE: The wc command counts UNIX line delimiters.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). use the UNIX command od –xc –Ax o To display the number of lines and characters in a specified ASCII text file. . 11. 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. transmitted. the generated OSH tab will appear in the Job Properties dialog box: Parallel Framework Red Book: Data Flow Job Design July 17. jobs are compiled into OSH (Orchestrate SHell) scripts that are used to execute the given job design at runtime. All rights reserved. 2006 126 of 179 © 2006 IBM Information Integration Solutions.

Sort (tsort) and Peek. 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. 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. Also available is some information about where data may be buffered. where data is repartitioned and how it is repartitioned. that is. transcribed. have been inserted by EE.5 Interpreting the Parallel Job Score When attempting to understand an Enterprise Edition flow. The job runs 3 processes on 2 nodes. It shows three stages: Generator. A score dump includes a variety of information about a flow. The following score dump shows a flow with a single Data Set. what degree of parallelism each operator runs with. which operators. . 2006 127 of 179 © 2006 IBM Information Integration Solutions. which has a hash partitioner that partitions on key field a. transmitted. ##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. if any. The Peek and Sort stages are combined.Information Integration Solutions Center of Excellence Figure 52: Generated OSH in Designer Job Properties 11. and exactly which nodes each operator runs on. including how composite operators and shared containers break down. they have been optimized into the same process. stored in a retrieval system.

Information Integration Solutions Center of Excellence lemond.p1] )} In a score dump.p0] lemond.com[op0. .torrent. transmitted. there are three areas to investigate: • Are there sequential stages? • Is needless repartitioning occurring? • In a cluster or Grid. 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. Parallel Framework Red Book: Data Flow Job Design July 17. are the computation-intensive stages shared evenly across all nodes? More details on interpreting the parallel job score can be found in 12. 2006 128 of 179 © 2006 IBM Information Integration Solutions.2 Understanding the Parallel Job Score.torrent. stored in a retrieval system. transcribed.com[op1.4.p0] )} op1[2p] {(parallel APT_CombinedOperatorController: (tsort) (peek) )on nodes ( lemond. All rights reserved.torrent.

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). Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. Performance tuning and optimization is an iterative process that begins at job design and unit tests. facilitating end-to-end parallelism across job boundaries. b) Remove unneeded columns as early as possible within the data flow. All rights reserved. proceeds through integration and volume testing.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. a) Use Parallel Data Sets to land intermediate result between parallel jobs. There are no utilities for examining data within a Lookup File Set. This section provides tips for designing a job for optimal performance. stored in a retrieval system. 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. However. File Sets facilitate parallel I/O at the expense of exporting to a specified file format. • Parallel Data Sets retain data partitioning and sort order. No part of this publication may be reproduced. Lookup File Sets can only be used on reference links to a Lookup stage. They maintain reference data in DS/EE internal format.1 How to Design a Job for Optimal Performance Overall job design can be the most significant factor in data flow performance. in the DS/EE native internal format. transmitted. or translated into any language in any form by any means without the written permission of IBM. If you need to share information with external applications. 2006 129 of 179 © 2006 IBM Information Integration Solutions. • Lookup File Sets can be used to store reference data used in subsequent jobs. and for optimizing the performance of a given data flow using various settings and features within DataStage Enterprise Edition. • Data Sets can only be read by other DS/EE parallel jobs (or the orchadmin command line utility). • When reading from database sources. pre-indexed. and continues throughout an application’s production lifecycle. This section outlines performance-related tips that can be followed when building a parallel data flow using DataStage Enterprise Edition. . 12.

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

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

It is for this reason that combination can be enabled or disabled on a per-stage basis.k. No part of this publication may be reproduced. Combined Operator Controller).Information Integration Solutions Center of Excellence However. Enterprise Edition is ‘greedy’ . 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. or translated into any language in any form by any means without the written permission of IBM. o Within Designer. tuning combination might yield additional performance benefits.a. when other performance tuning measures have been applied and still greater performance is needed. 2006 132 of 179 © 2006 IBM Information Integration Solutions.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’. There are 2 ways to affect operator combination: o The environment variable APT_DISABLE_COMBINATION. transmitted. this is only recommended on pre 7. the assumptions used by the Enterprise Edition optimizer to determine which stages can be combined may not always be the most efficient. (For information on interpreting a job score dump.) In addition. or globally.0 versions of DS/EE. However. combination can be set on a per-stage basis (on the Stage/Advanced tab) The job score identifies what components are combined. combined stages are indicated by parenthesis surrounding the % CPU.4.2 Understanding the Parallel Job Score” in this document. stored in a retrieval system. disables ALL combination in the entire data flow. transcribed. All rights reserved. see 12. 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. it is best to let DSEE decide what to combine and what to leave uncombined. if the “%CPU” column is displayed in a Job Monitor window in Director. . When deciding which operators to include in a particular combined operator (a.

DS/EE executes a given job across the resources defined in a the specified configuration file. 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. This will prevent IO requests from waiting on CPU to become available and viceversa. as shown in this subsequent Job Monitor for the same job: Figure 54: Throughput in Job Monitor after disabling combination 12. Lookup. There are times. if you have several transformers and database operators combined with an output Sequential File. In fact.) from CPU-heavy operators (Transformer.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. etc. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. 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. Change Capture. No part of this publication may be reproduced. However. it is good to separate I/O heavy operators (Sequential File. such as Join. transcribed. it might be a good idea to set the sequential file to be non-combinable.). etc. however. the I/O-intensive FileSet is combined with a CPU-intensive Transformer. Disabling combination with the Transformer enables pipeline partitioning. in the above job design. Merge. . and improves performance. a single-node configuration file is often appropriate to minimize job startup time and resource requirements without significantly impacting overall performance. 2006 133 of 179 © 2006 IBM Information Integration Solutions. For example. when it is appropriate to minimize the resource requirements for a given scenario.Information Integration Solutions Center of Excellence Choosing which operators to disallow combination for is as much art as science. Funnel) Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. stored in a retrieval system. in general. Full Sorts. Toward that end.

then setting APT_DEFAULT_TRANSPORT_BLOCK_SIZE to a multiple of (or equal to) the record size will resolve the problem. the value provided by a user for this variable is rounded up to the operating system's nearest page size. Remember. This type of buffering (or ‘Record Blocking’) is rarely tuned. or translated into any language in any form by any means without the written permission of IBM. there are 2 of these transport blocks for each partition of each link.1 Inter-Operator Transport Buffering Though it may look like it from the performance statistics and documentation might discuss ‘record streaming’. They are actually transferred in blocks (just like on old magnetic tapes) called “Transport Blocks”. The second block will be used by the downstream/consumer stage to obtain data that is ready for the next processing step. stored in a retrieval system. transcribed. All rights reserved. ODBC Enterprise. strictly speaking. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MAX_TRANSPORT_BLOCK_SIZE APT_MAX_TRANSPORT_BLOCK_SIZE July 17. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. Funnel. Cannot be less than 8192.) 12.4 Understanding Buffering There are two types of buffering in Enterprise Edition: ‘inter-operator transport’ and ‘deadlock prevention’. Merge. BufferOp Selectively (being careful to to avoid deadlocks) disable buffering. records do not stream from one stage to another. Aggregator. cannot be greater than 1048576.Information Integration Solutions Center of Excellence   Minimize non-combinable stages (as outlined in the previous section) such as Join. . The first block will be used by the upstream/producer stage to output data it is done with. It usually only comes into play when the size of a single record exceeds the default size of the transport block. BuildOps. 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. (Buffering is discussed in more detail in the following section. 12. 2006 134 of 179   Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Default is 8192. with a valid value range for between 8192 and 1048576. so setting this value too high will result in a large amount of memory consumption. If necessary.4. transmitted. Oracle Enterprise. Each pair of operators that have a producer/consumer relationship will share at least 2 of these blocks. the blocks are swapped and the process begins again. APT_MIN_TRANSPORT_BLOCK_SIZE o Specifies the minimum allowable block size for transferring data between players. Once the upstream block is full and the downstream block is empty. DB2 Enterprise. No part of this publication may be reproduced. Remove Duplicates. The default value is 8192.

APT_LATENCY_COEFFICIENT. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. 12.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players. 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.2 Deadlock Prevention Buffering The other type of buffering. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE. transcribed. No part of this publication may be reproduced. Default is 1048576. stored in a retrieval system. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records.4. Here is an example job fragment: Figure 55: Fork-Join example 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. Cannot be less than 8192. transmitted. cannot be greater than 1048576. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. 2006 135 of 179 © 2006 IBM Information Integration Solutions. “Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job. . All rights reserved.

and Aggregator2 is waiting on Transformer. Aggregator1 is waiting to write to the Join. or translated into any language in any form by any means without the written permission of IBM. the job would deadlock . d eue Qu ad Re Transformer Waiting to write to Aggregator1 Qu eue Re d ad Join Waiting to read from Aggregator2 July 17. BufferOp is always ready to read or write and will not allow a read/write request to be queued. there is a specialized operator called BufferOp. it doesn’t necessarily have to involve a Join stage. No part of this publication may be reproduced. Imagine that the Transformer is waiting to write to Aggregator1. To understand deadlock-prevention buffering. and Join is waiting to read from Aggregator2. Without deadlock buffering. it would eventually time out). To guarantee that this problem never happens in Enterprise Edition. the Transformer creates a fork with 2 parallel Aggregators. Note however.Information Integration Solutions Center of Excellence In this example. transmitted.) Without deadlock buffering. transcribed. 2006 136 of 179 . instead of data flow. Aggregator1 is waiting on Join. stored in a retrieval system. All rights reserved. Aggregator2 is waiting to read from the Transformer. 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. 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. it is important to understand the operation of a parallel pipeline. which go into an Inner Join. 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.bringing processing to a halt (though the job does not stop running. Join is waiting on Aggregator2. It is placed on all inputs to a join structure (again.

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. No part of this publication may be reproduced. “The Parallel Engine Configuration File” of the DataStage Manager guide).) Since BufferOp is always ready to read or write. as these same types of circular dependencies can result from partition-wise Fork-Joins. BufferOps will allocate 3MB of memory each (remember that this is per operator. Values greater than 1 indicate that the buffer July 17. Join cannot be ‘stuck’ waiting to read from either of its inputs. stored in a retrieval system. TIP: For very wide rows. 2006 137 of 179  Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. All rights reserved. . it may be necessary to increase the default buffer size (APT_BUFFER_MAXIMUM_MEMORY) to hold more rows in memory. or translated into any language in any form by any means without the written permission of IBM. 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. transcribed. per partition). APT_BUFFER_FREE_RUN o Maximum capacity of the buffer operator before it starts to offer resistance to incoming flow. By default.Information Integration Solutions Center of Excellence BufferOp1 Aggregator 1 Transformer Join Aggregator2 BufferOp 2 (Here the arrows now represent data-flow. transmitted. not dependency. Default 1 MB. thus breaking the circular dependency and guaranteeing no deadlock will occur. May not exceed 2/3 of APT_BUFFER_MAXIMUM_MEMORY. 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.

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

Choosing which stages to tune buffering for and which to leave alone is as much art as science. No part of this publication may be reproduced. . when a buffer has consumed its RAM. All rights reserved. as some other stage far downstream could be responsible for cascading pushback all the way upstream to the place you are seeing the bottleneck. you cannot determine that any one stage is waiting on any other stage. Because of this. then setting the buffering policy to “FORCE_BUFFERING” will cause buffering to occur everywhere. Stages upstream/downstream from high-latency stages (such as remote databases.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. while changing only 1 thing at a time).) are a good place to start. stored in a retrieval system. By using the performance statistics in conjunction with this buffering. As implied above. transmitted. it will ask the upstream stage to “slow down” . you may be able identify points in the data flow where a downstream stage is waiting on an upstream stage to produce data. and should be considered among the last resorts for performance tuning. When attempting to address these mismatches in production/consumption.this is called “pushback”. Each place may offer an opportunity for buffer tuning. If that doesn’t yield enough of a performance boost (remember to test iteratively. instead of globally through environment variable settings. etc. 2006 139 of 179 © 2006 IBM Information Integration Solutions. transcribed. it is best to tune the buffers on a per-stage basis. if you do not have force buffering set and APT_BUFFER_FREE_RUN set to at least ~1000. 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. NFS mount points for data storage.

2006 140 of 179 © 2006 IBM Information Integration Solutions. stage names. using the Manager DSX export capability. .  Create a standard directory structure outside of the DataStage project directory for source and target files. An example DataStage naming structure is given in Section 2.3: Documentation and Annotation. To ease re-use. source and target file paths.Information Integration Solutions Center of Excellence Appendix A: Standard Practices Summary This Appendix summarizes Standard Practices recommendations outlined in this document.5: Job Parameters. database login settings. or translated into any language in any form by any means without the written permission of IBM. Development Guidelines Modular development techniques should be used to maximize re-use of DataStage jobs and components.2: Naming Conventions.1 Environment Variable Settings)  annotation blocks Job Parameters should always be used for file paths. 2. especially for DataStage Project categories.  Where possible.1:Directory Structures. database login properties…)  environment variables and their default settings (as outlined in Section 2. See Section 2. It is the DataStage developer’s responsibility to make personal backups of their work on their local workstation. etc) and directories where files are stored. these standard parameters and settings should be made part of a Designer Job Template. The scope of a parameter is discussed further in Section 3.  Naming conventions. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. intermediate work files. as well as Annotation fields. along with cross-references for more detail. password. This can also be used for integration with source code control systems. 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. and links. 1. create re-usable components such as parallel shared containers to encapsulate frequently-used logic. All rights reserved. All DataStage jobs should be documented with Short Description field. An example directory naming structure is given in Section 2. No part of this publication may be reproduced. user. DataStage Template jobs should be created with:  standard parameters (for example. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. and so forth. stored in a retrieval system. Standards It is important to establish and follow consistent standards in:  Directory structures for install and application support directories.5.4: Working with Source Code Control Systems.  A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN. transmitted. file names. See Section 2.

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

All rights reserved. use Auto partitioning (the default) Parallel Framework Red Book: Data Flow Job Design July 17. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. persistent Data Sets can be used to retain the partitioning and sort order. stored in a retrieval system. but is not limited to: Aggregator. It may also be necessary for Transformers and BuildOps that process groups of related records. Using the above objectives as a guide. 2006 142 of 179 © 2006 IBM Information Integration Solutions. minimize repartitioning. Remove Duplicates. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. Change Apply. Merge. or translated into any language in any form by any means without the written permission of IBM. Change Capture. transmitted. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). No part of this publication may be reproduced. 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. . examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. 6. 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. 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. Note that in satisfying the requirements of this second objective. 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. and Sort stages. Collecting Data Given the options for collecting data into a sequential stream. 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. transcribed. Join. Across jobs.Information Integration Solutions Center of Excellence This includes.

or translated into any language in any form by any means without the written permission of IBM. o Always place a reject link on a parallel Transformer to capture / audit possible rejects. sequential ordered result set use a parallel Sort and a Sort Merge collector 8. 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. use Sort Merge collector to produce a single. transmitted. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. as long as the Data Set has not been repartitioned or reduced. use the Join or Merge stage. Join vs. See Section 9. 2006 143 of 179 © 2006 IBM Information Integration Solutions. transcribed. 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. Create Key Change Column. stored in a retrieval system.Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel.  Stage-Specific Guidelines As discussed in Section 8. The Lookup stage is most appropriate when reference data is small enough to fit into available memory. 7. Merge. All rights reserved.1. Sorting Using the rules and behavior outlined in Section 6: Sorting. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. Create Cluster Key Change Column. 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.1: Lookup vs. 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.1: Transformer NULL Handling and Reject Link.  Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. . If the Data Sets are larger than available memory resources.

All rights reserved. A Sort method Aggregator should be used when the number of distinct key values is large or unknown.  Database Stage Guidelines Where possible. the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records. use a SQL where clause to limit the number of rows sent to a DataStage job. or when exception processing. as discussed in Section 10. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. it is best to implement business rules natively using DataStage parallel components. transcribed. 2006 144 of 179 © 2006 IBM Information Integration Solutions.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. use orchdbutil to properly import design metadata. transmitted. Use Hash method Aggregators only when the number of distinct key column values is small. For maximum scalability and parallel performance. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. 10. When using Oracle. No part of this publication may be reproduced. If possible.2Capturing Unmatched Records from a Join). which may indicate an underlying problem or data type conversion issue.1. 9. Check the Director log for warnings. or Informix databases. DB2.1: 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. 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. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. Parallel Framework Red Book: Data Flow Job Design July 17. (See Section 9. 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 the native parallel database stages for maximum performance and scalability. .

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

Subject Modifier. 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. 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. . File.Information Integration Solutions Center of Excellence Appendix B: DataStage Naming Reference Every name should be based on a three-part concept: Subject. No part of this publication may be reproduced. transmitted. Class Word where the following frequently-used Class Words describe the object type. stored in a retrieval system. transcribed. 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. All rights reserved. or translated into any language in any form by any means without the written permission of IBM.

transcribed. . All rights reserved. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. 2006 147 of 179 © 2006 IBM Information Integration Solutions. transmitted.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. No part of this publication may be reproduced.

Viewing the Job Score When the environment variable APT_DUMP_SCORE is set. as the job score offers invaluable data for debugging and performance tuning. All rights reserved. and the overhead to capture the score is negligible. and can be ignored. multiple operators are combined within a single operating system process to improve performance and optimize resource requirements. The second score entry is the actual job score. Enterprise Edition uses the given job design and configuration file to compose a job score which details the processes created. transmitted. Similar to the way a parallel database optimizer builds a query plan. . At runtime. Parallel Framework Red Book: Data Flow Job Design July 17. 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. not the actual job. transcribed. The first score is from the license operator. As shown in the illustration below. 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. degree of parallelism and node (server) assignments. or translated into any language in any form by any means without the written permission of IBM. and interconnects (Data Sets) between them. It is recommended that this setting be enabled by default at the project level. 2006 148 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. 1. the job score is output to the DataStage Director log. 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.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. stored in a retrieval system.

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

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

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

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

add (# nullable fields)/8 for null indicators . Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. 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: . . 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.Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. 2006 153 of 179 © 2006 IBM Information Integration Solutions.one byte per column for field alignment (worst case is 3.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries. or translated into any language in any form by any means without the written permission of IBM. if you have a transfer that transfers all fields from input to output. the method APT_Record::estimateFinalOutputSize() can give you an estimate for a given record schema. transcribed. No part of this publication may be reproduced. transmitted. As can APT_Transfer::getTransferBufferSize(). All rights reserved.

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

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

Useful in debugging Oracle SQL*Loader issues. By default. When set. Commits are made whenever the time interval period has passed or the row interval is reached. Specifies Oracle SQL*Loader options used in a target Oracle stage with Load method. Normally set in a user’s environment by Oracle scripts. this is set to OPTIONS(DIRECT=TRUE. Informix Environment Variables Environment Variable $INFORMIXDIR $INFORMIXSQLHOSTS $INFORMIXSERVER $APT_COMMIT_INTERVAL Setting [path] [filepath] [name] [rows] Description Specifies the Informix install directory. No part of this publication may be reproduced. The default is 10000 per partiton. transcribed. Environment Variable $ORACLE_HOME Oracle Environment Variables Setting [path] [sid] [num] [seconds] Description Specifies installation directory for current Oracle instance.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. stored in a retrieval system. . Specifies the Oracle service name. corresponding to a TNSNAMES entry. All rights reserved. a target Oracle stage with Load method will limit the number of players to the number of datafiles in the table’s tablespace. Specifies the path to the Informix sqlhosts file. 5. July 17.Information Integration Solutions Center of Excellence 4. 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. Setting this variable makes it possible to load fields with trailing or leading blank characters. The filepath specified by this environment variable specifies the file with the SQL*Loader commands. Specifies the name of the Informix server matching an entry in the sqlhosts file. 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. These two environment variables work together to specify how often target rows are committed for target Oracle stages with Upsert method. whichever comes first. By default. When set. 2006 156 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. the output of a Target Oracle stage with Load method is written to files instead of invoking the Oracle SQL*Loader. Specifies the commit interval in rows for Informix HPL Loads. transmitted.

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.

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

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

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. transmitted. Here is another possible outcome: Also: Consider the result of running the same job with the same data. . but a different number of partitions. stored in a retrieval system. July 17. Minimizing Record Movement for Maximizing Performance Now we have covered the basic rules and mechanics for hash-partitioning and sorting. 6 There is an exception to this rule: If your hash key has only one value. 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.Information Integration Solutions Center of Excellence 3 102 Example: 6 rows in 2 partitions. transcribed. No part of this publication may be reproduced. All rights reserved.

2006 163 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 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. 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. per store for all stores in the nation. 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. and append these values to the original data. They wish to determine the weighted average transaction amount peritem nation-wide. All rights reserved. as well as the average transaction amount per-item. . transcribed. 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. 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. thereby increasing the value of this exercise.Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets nation-wide.

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

5 180.5 3. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’.5 Store Average Item Transaction Amt 28 28 43 43 6. 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. transmitted.5 46 46 6.25 26. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation. 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. It does this by keeping running totals in memory for the aggregation for each output group.5 6. The sort method requires the input data to be hashed and sorted. All rights reserved.5 3.5 16.5 310. it does not guarantee output order. in return. or translated into any language in any form by any means without the written permission of IBM.25 26.5 50. The hash method only requires the input data to be hashed.5 6. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.5 35. however. you will need to set the Aggregator’s “Method” to Sort.25 180.5 35. .25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17.5 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. 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.5 310.5 50.25 26.5 35.25 16.5 39.5 39. Therefore. transcribed.5 23 23 23 23 26.5 180. not Hash. No part of this publication may be reproduced.5 180.Information Integration Solutions Center of Excellence NOTE: For this scenario.5 3. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35. stored in a retrieval system.

as only partition concurrency is affected by sequential execution. and 1.25 44. A much better answer is: 1. i.25 44. Parallel Framework Red Book: Data Flow Job Design July 17.75 38.5 33. stored in a retrieval system. and 6.25 44. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition. This is certainly an improvement on the previous answer.5 61. transmitted. .25 16. more efficientv. An examination of the job above would suggest: 6. All rights reserved. and 3. 3.e. No part of this publication may be reproduced. transcribed. 6. appended to the end of this document for masochistsiv) might suggest: 4.5 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.25 44.: record adjacency assumes partition concurrency8. A deeper examination (of the score dump.75 38. 2006 166 of 179 © 2006 IBM Information Integration Solutions.75 26.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. even sequential job execution does not alter the answer for the sort-related questions. or translated into any language in any form by any means without the written permission of IBM. Here’s a screen shot of this.5 27 27 61.25 38. 3.5 33.75 38.

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

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

Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 44. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17.5 33.5 33.75 38. Output from above solution: Data Set 3: PeekFinalOutput.25 44.25 38.5 27 27 61. All rights reserved.75 Store Average Item Transaction Amt 2 2 6 6 5 5 26.5 61. No part of this publication may be reproduced. and it is enabled by default. 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.5 3. useful for preserving previous sort orderings).25 44.25 3. transmitted.5 16.Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt. stored in a retrieval system. as it is much more expensive than non-stable sorts.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. or translated into any language in any form by any means without the written permission of IBM. 2006 169 of 179 © 2006 IBM Information Integration Solutions.25 16. transcribed.75 38.75 38. .25 16.5 26.5 3.25 44.5 16.

5 310.000.25 26.5 35.25 26. a savings of 400. but there is a critical difference.000.25 26. transcribed.5 180.5 310.000.5 39.000 records.25 180. stored in a retrieval system.5 23 23 26.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. 000.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume.5 This solution produces the same result but is achieved with only one complete sort. That is a LOT of saved processing power.000 records.000 records in addition to hashing 300. transmitted. The second solution only sorts (on disk) 100. and only hashes.5 35.5 180.5 50. All rights reserved. we had to sort (on disk) 300. Parallel Framework Red Book: Data Flow Job Design July 17. .5 35.5 6. and two sub-sorts—a much more efficient solution for large data volumes. or translated into any language in any form by any means without the written permission of IBM.5 50.5 39. It looks very similar to the first solution.000 records. No part of this publication may be reproduced.5 6. 2006 170 of 179 © 2006 IBM Information Integration Solutions. a single partitioner. With the initial solution. 100.5 Store Average Item Transaction 28 28 43 43 6.5 23 23 180. There is an even more efficient solution.5 46 46 6.000. Imagine a job with 100 million records as the input.

stored in a retrieval system.5 16. 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. This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator. No part of this publication may be reproduced.25 44.a significant savings. Data Set 4: PeekFinalOutput. However it will not create additional processes to handle the records and re-order them. Here is the output from this version of the job. transmitted. Also.5 26.5 3. All rights reserved. Comparing the efficiency of this solution with that of number two. Looks a lot like solution 1.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26.25 44.Information Integration Solutions Center of Excellence NOTE: In this job.25 16.25 16. or translated into any language in any form by any means without the written permission of IBM.5 3. we saved a sub-sort on 100 million records . we have chosen to use the StoreLocation column as a part of our sorting key. and automatic partition insertion must be turned off vii. for the same reasons. automatic sort insertion. but NOT to use it for hashing. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.5 3. transcribed.25 16. This is a potentially huge savings on large data volumes (remember the previous example). the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed. except w/o the sort on the output of JoinSourceToAggregator. The difference is on DistributeCopiesOfSourceData: Here.5 27 27 July 17. .

5 35.25 26. 2006 172 of 179 © 2006 IBM Information Integration Solutions. in addition to the heavy penalty paid in disk I/O for using a full sort.5 44 44 PeekFinalOutput.5 46 46 6.25 26.5 180. or translated into any language in any form by any means without the written permission of IBM.5 180.25 26.5 180.25 44.25 180.5 39.5 310. stored in a retrieval system. All rights reserved. transcribed.5 50.5 39.5 50.25 38. sort. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 33.5 Store Average Item Transaction Amt 28 28 43 43 6.5 23 23 23 23 26.75 38. transmitted.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. by definition. No part of this publication may be reproduced.75 38.5 6.5 35.75 61.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. .5 310.5 33.75 38. 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 35. Here is a screen shot of a sort running on 40 million records: Parallel Framework Red Book: Data Flow Job Design July 17.5 6.

. 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. This allows down-stream stages to be processing data during the sorting process. This is because a standard sort requires all rows to be present in order to release the first row. All rights reserved. the link sort in solution 2. no rows have left yet. 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). requiring a large amount of scratch diskx. transmitted. we are sub-sorting the data we sorted in the previous diagram). although ~5 million records have entered the sort. Here is an example of a ‘sub-sort’: Here. No part of this publication may be reproduced. This situation is analogous to all of the sorts in solution 1. 2006 173 of 179 © 2006 IBM Information Integration Solutions. instead of waiting until all 40 million records have been sorted (in this instance. and the link sort in solution three. stored in a retrieval system. transcribed.Information Integration Solutions Center of Excellence As you can see.

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

p1] 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] )} op6[3p] {(parallel buffer(0)) .p1] node3[op4.DSLink2) on nodes ( node1[op0.p0] node2[op3. subArgs={ cs } } })#>eCollectAny op10[3p] (parallel JoinSourceToAggregator_2.p0] node2[op2.p1] node1[op1.p1] node3[op5. key={ value=TransactionDate }.p0] node1[op1.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op1.p0] node2[op5.p2] )} op2[3p] {(parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op2.p1] node3[op2. key={ value=StoreLocation.p2] )} op5[3p] {(parallel APT_TSortOperator(1)) on nodes ( node1[op5.p0] node1[op0. subArgs={ cs } }.p2] )} op3[3p] {(parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op3.DSLink18_Sort)} ds12: {op10[3p] (parallel JoinSourceToAggregator_2.p0] node2[op4.[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.p2] )} op4[3p] {(parallel APT_TSortOperator(0)) on nodes ( node1[op4.p1] node3[op3.

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

p2] )} op4[3p] {(parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op4.p1] node3[op1.[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] node1[op0.p2] .p0] node2[op1.p1] node3[op3.DSLink2_Sort) on nodes ( node1[op1.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta.p0] node1[op0.p1] node3[op4.p1] node3[op2.p0] node2[op5.p2] )} op3[3p] {(parallel SubSortOnStoreLocation) on nodes ( node1[op3.DSLink2) on nodes ( node1[op0.p0] node2[op3.p1] node3[op5.p0] node2[op2.p0] node2[op4.p2] )} op2[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op2.p2] )} op5[3p] {(parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op5.

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

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

Sign up to vote on this title
UsefulNot useful