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.

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

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

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

built from individual Parallel jobs assembled in DataStage Enterprise Edition (“DS/EE”) controlled as modules from master DataStage Sequence jobs. etc). etc). . These topics are discussed in Parallel Framework Standard Practices: Administration. No part of this publication may be reproduced. as illustrated in the example below: These job Sequences control the interaction and error handling between individual DataStage jobs.Information Integration Solutions Center of Excellence 1.1 Job Sequences As shown in the previous diagram. or translated into any language in any form by any means without the written permission of IBM. Job Sequences require a level of integration with various production automation technologies (scheduling. In most production deployments. Job sequences also provide the recommended level of integration with external schedulers (such as AutoSys. and together form a single end-to-end module within a DataStage application. error logging. Cron. CA7. 2006 7 of 179 © 2006 IBM Information Integration Solutions. ETL development is intended to be modular. Management. Parallel Framework Red Book: Data Flow Job Design July 17. and provides an appropriate leveraging of the respective technologies. transcribed. This provides a level of granularity and control that is easy to manage and maintain. All rights reserved. and Production Automation. stored in a retrieval system. auditing/capture. transmitted.

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

The following example transformation job demonstrates the use of write-through cache DS/EE Data Sets: The target table is among the inputs.2. that Data Set qualifies as write-through cache and may be used as source data instead of the target table. 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. All rights reserved. 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. No part of this publication may be reproduced. transcribed. some of which may be write-through cache Data Sets.Information Integration Solutions Center of Excellence 1. stored in a retrieval system. 2006 9 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17.1 Transformation Jobs In transformation jobs. data sources. . transmitted.

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

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

2. 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.3 Provisioning Jobs This example provisioning job demonstrates the straightforward approach to simple provisioning tasks. transmitted. 2006 12 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence 1. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. stored in a retrieval system. . All rights reserved.

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

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

transmitted. /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. /Project_Z Within the separate Staging file system. .. 2006 15 of 179 . /Staging /dev Top-Level Directory development data tree. files are separated by Project name as shown below./archive of these development phases may be present on a local file not all system.. /Project_Z July 17.. No part of this publication may be reproduced..Information Integration Solutions Center of Excellence /Staging /dev /si /qa /prod /Project_A /Project_A /Project_A /Project_A /archive /archive /archive /archive . 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 . Within each deployment directory. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system.. data directories are implemented for each deployment phase of a job (development... location of source data files. If the file system is not /archive /archive /archive shared across multiple servers. transcribed.. system integration. and production) as appropriate. All rights reserved. qa. . target data files. error and reject files.

which will enable DataStage applications to move through the life cycle without any code changes. . A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project... July 17.. This will provide a convenient vehicle to group and manage resources used by a project.Information Integration Solutions Center of Excellence 2.. such as the operating system. or it could require scripts for example integrating with an Enterprise Scheduler..1. transmitted. 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. transcribed. another application or middle ware. All rights reserved.. The integration can be as simple as a file system for housing source files. . stored in a No part of this /Project_Z /Project_Z retrieval system. /Project_Z . The directory structure will be made transparent to the DataStage application. through the use of environment variables. 2006 .. Environment variables are a critical portability tool. or translated into any language in any form by any means without the written permission of IBM. 16 of 179 .2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities.. 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./Project_Z publication may be reproduced.

utilities. Parallel Framework Red Book: Data Flow Job Design July 17. directories are created for each deployment phase of a job (development. system integration.ds file) location of custom job logs and reports location of parameter files for automated program control. If the file system is not shared across multiple servers. No part of this publication may be reproduced. for example the following diagram shows Project_Plus variables as defined in the DataStage Administrator. not all of these development phases may be present on a local file system. qa. BuildOps. a copy of dsenv and copies of DSParams. stored in a retrieval system. DataStage routines. . 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. 2006 17 of 179 © 2006 IBM Information Integration Solutions. transmitted. or translated into any language in any form by any means without the written permission of IBM. transcribed. All rights reserved. and production) as appropriate.Information Integration Solutions Center of Excellence Within the Project_Plus hierarchy.$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. 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. this directory may only be present in the /dev development code tree) location of DataSet header files (.

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

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

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

For example.4.1 Projects Each DataStage Project is a standalone repository.4. No part of this publication may be reproduced. transcribed. Routines.2. Table definitions…). 2006 21 of 179 © 2006 IBM Information Integration Solutions. and Production (“Prod”). Category Names can be long. This factor often can cause terminology issues especially in teamwork where both business and developers are involved. Therefore Directory names should be Word Capitalized and separated by either an underscore or a space. it can contain alpha-numeric characters and it can contain underscores.4. Below this level.4 Naming Conventions by Object Type 2. It may or may not have a one to one relationship with an organizations’ project of work. The name of a DataStage Project may only be 18 characters in length. DataStage enforces the top level Directory Structure for different types of Objects (for example.3 Job Category Naming Within Designer. transmitted.2.2. Shared Containers. The suffix of a Project name should be used to identify Development (“Dev”).2. allowing related objects to be grouped together. 2. developers have the flexibility to define their own Directory or Category hierarchy. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17. Test (“Test”). 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. which can provide impact analysis. Jobs.Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage. 2. stored in a retrieval system. as well as documentation and auditing. or translated into any language in any form by any means without the written permission of IBM. However with the limit of 18 characters the name is most often composed of abbreviations. dialog box fields that specify a new category have only one input area for defining the Category name.2 Category Hierarchy DataStage organizes objects in its repository by Categories. . are Alpha Numeric and can also contain both Spaces and Underscores. All rights reserved. Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”).

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

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

An alternative implementation is to set the “Data source name” to that of the source system or schema.4. with sub-categories intended to identify the type of the source. In this example. or translated into any language in any form by any means without the written permission of IBM. and underscores only. All rights reserved.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters. transmitted. No part of this publication may be reproduced. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category. the TableDefs have been grouped into a master category of Custom. numbers.g: DWPH1 or ECRP. Parallel Framework Red Book: Data Flow Job Design July 17. e. Figure 7: Table Definition Categories 2.2.2. This will be discussed further in Section 2. transcribed.: Datasets. . Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”.4. stored in a retrieval system. 2006 24 of 179 © 2006 IBM Information Integration Solutions.2 Category Hierarchy. e. Each subject area will have a master category. Because the name of can be long.Information Integration Solutions Center of Excellence should be deleted as soon as possible. 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. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”. Job and Job Sequence names should be descriptive and should use word capitalization to make them readable.g.

stage editors identify links by name.Information Integration Solutions Center of Excellence 2. Within the graphical Designer environment. To differentiate between Parallel Shared Containers and Server Shared Containers. a character code is automatically added to that instance of its use throughout the project. All rights reserved.2. transmitted. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2. during Link Ordering).6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. Within a Job Sequence. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. or translated into any language in any form by any means without the written permission of IBM. Furthermore. transcribed. No part of this publication may be reproduced.4.8 Links Within a DataStage Job. 2006 25 of 179 © 2006 IBM Information Integration Solutions.2. stored in a retrieval system. 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. It is particularly important to establish a consistent naming convention for link names.4.2. The class word suffix is “Parm”.7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores. . 2. instead of using the default “DSLink#” (where “#” is an assigned number). 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. links are objects that represent the flow of data from one stage to the next. links represent the flow of a message from one activity / step to the next.4. having a descriptive link name reduces the chance for errors (for example. When a Shared Container is used. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. It is optional as to whether you decide to change this code to something meaningful. Therefore the parameter name must be made readable using Capitalized words separated by underscores. so word capitalization should be used for readability.

through Job reporting). based on the order the object was added to the flow. establishing standardized link names makes it easier to understand results and audit counts.4.Information Integration Solutions Center of Excellence example. 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.2 DataStage Naming Reference. Transformer. etc) • The type of movement may optionally be part of the Class Word. Within a Job or Job Sequence. stage names must be unique. Database.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. reject) to reinforce the visual cues of the Designer canvas: o “Ref” for reference links (Lookup) o “Rej” for reject links (Lookup. stored in a retrieval system. In this case. 3. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12. 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. Sequential File. These names are based on the type of stage (object) and a unique number. Parallel Framework Red Book: Data Flow Job Design July 17. Merge. transcribed. or 4 character abbreviation should be used for the Class Word suffix.4.2. All rights reserved. (The DataStage Designer does not require link names on different stages to be unique. always specify a unique link name within a particular Job or Job Sequence by including a number. after the subject name and subject modifier. . the same name may be appropriate for multiple links.) • 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. transmitted. No part of this publication may be reproduced. 2006 26 of 179 © 2006 IBM Information Integration Solutions. a 2. Instead of using the full object name. the link name should include the link type (reference. or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

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

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

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

transmitted. No part of this publication may be reproduced. or by the individual DataStage developer. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu. 2. stored in a retrieval system.2 Category Hierarchy. and individual Object levels. The DataStage Manager can export at the Project. the developer should create a local backup prior to implementing any extensive changes. All rights reserved. The DataStage client includes Windows command-line utilities for automating the export process. There is no server-side project export facility.4. Category.4. establishing and following a consistent naming and categorization standard is essential to the change management 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. . 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. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. 2. • Select the object or category in the Manager browser.2 Using Object Categorization Standards As discussed in Section 2.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process.2. These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide.Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group. as explained in this section. transcribed. July 17.4. 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.DSX file 2. check-in the exported . 2006 34 of 179 © 2006 IBM Information Integration Solutions. Figure 11: Manager Export Options To export a group of objects to a single export file. the import of objects from an external source code control system is a straightforward process. or translated into any language in any form by any means without the written permission of IBM. make sure the “Job Executables” category is checked. . Parallel Framework Red Book: Data Flow Job Design July 17. the option “Selection: By category” should be specified in the “Options” tab. No part of this publication may be reproduced.4. transmitted. • Using your source code control utilities. stored in a retrieval system. or automated through command-line utilities. transcribed. Make sure all objects are saved and closed before exporting.Information Integration Solutions Center of Excellence NOTE: Objects cannot be exported from DataStage if they are open in Designer. Import can be interactive through the Manager client (as described in this section). If you wish to include compiled Transformer objects for a selected job. All rights reserved.4 Import from Source Code Control System In a similar manner.

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

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

but part of every job design so that it can be easily enabled for debugging purposes. . transcribed. $UNSET 2. 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. or translated into any language in any form by any means without the written permission of IBM. APT_SORT_INSERTION_CHECK_ONLY). This variable should be included in all job parameters so that it can be easily changed at runtime. The count is per operator per partition. This section is intended to call attention to some specific environment variables.5.1. Several environment variables are evaluated only for their presence in the environment (for example. processes. This setting should be disabled by default.5. This setting should be disabled by default.4. but part of every job design so that it can be easily enabled for debugging purposes. stored in a retrieval system.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment. It may be helpful to create a Job Template and include these environment variables in the parameter settings. (see 12. Does not report PIDs of DataStage “phantom” processes started by Server shared containers. or may be set on an individual basis within the properties for each job. 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. These settings can be made at the project level.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. Extremely useful for understanding how a job actually ran in the environment. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. transmitted. providing detailed information about actual job flow including operators. and Data Sets. Outputs EE score dump to the DataStage job log.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. All rights reserved. 2006 37 of 179 © 2006 IBM Information Integration Solutions.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. and to document a few that are not part of the documentation. Outputs actual runtime metadata (schema) to DataStage job log. No part of this publication may be reproduced. 2.

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

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

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.

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

Pass successful reads to the output stream. 2006 44 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM.rej file. Only records that match the given table definition and format are output. Push back – rows are sent to a Data Steward for corrective action.1 Reject Handling with the Sequential File Stage The Sequential File stage can optionally include a reject link. Alternatively. transmitted. These actions include: 1. and should be reviewed by the Data Steward. The default action is to push back reject and error rows to a Data Steward. Pass successful reads to the output stream. All rights reserved. 2 3 Rejects should not exist but should not stop the job. 3. which outputs rows that do not match the given table definition and format specifications. 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. Reprocess – rows are reprocessed and re-enter the data stream. (No reject link exists) Fail Abort job on read format failure (No reject link exists) Output Reject switch failures to the reject stream. 2. Ignore – some process or event downstream of the ETL process is responsible for handling the error. Send the reject stream to a *. By default. Use the Fail option. No part of this publication may be reproduced.7. (Reject link exists) The reject option should be used in all cases where active management of the rejects is required. it must have a *. A message is always written to the Director log which details the count of rows successfully read and rows rejected. stored in a retrieval system.rej file extension. Rejects should not exist and should stop the job. . transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. a shared container error handler can be used. Rejects are tracked by count only. If a file is created by this option. rows that cannot be read are dropped by the Sequential File stage.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. Use the Output option. 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. .Information Integration Solutions Center of Excellence 3. All rights reserved. each must be done in its own Lookup. Pass successful lookups to the output stream.rej extension is used when rejects require investigation after a job run. Lookup Stage Option Drop if lookup fields are necessary down stream or Continue if lookup fields are optional Send the reject stream to an *. it must have a *. 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. Use the Fail option.7.rej file or tag and merge with the output stream.2 Reject Handling with the Lookup Stage The Lookup stage compares a single input stream to one or more reference streams using keys. Drop Drop lookup failures from the input stream. If a file is created by this option. transcribed. The reject option should be used in all cases where active management of the rejects is required.7. If there are multiple validations to perform. Fail Abort job on lookup failure Reject Reject lookup failures to the reject stream. stored in a retrieval system.err file or tag and merge with the output stream. 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. however. to enforce error management ONLY ONE REFERENCE LINK is allowed on a Lookup stage. This behavior makes the Lookup stage very valuable for positive (reference is found) and negative (reference is NOT found) business rule validation. or translated into any language in any form by any means without the written permission of IBM. a local error handler based on a shared container can be used.err file extension. Rejects should not exist but should not stop the job. Pass successful lookups to the output stream. transmitted. Pass successful lookups to the output stream. Send the reject stream to an *.err extension when rejects can be ignored but need to be recorded. and rejects can occur if the key fields are not found in the reference data. and should be reviewed by the Data Steward. Furthermore. they only need to be recorded but not acted on. 2 3 4 3. Alternatively. 2006 45 of 179 © 2006 IBM Information Integration Solutions. Rejects should not exist and should stop the job.rej or *. No part of this publication may be reproduced. the *. The *.

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

This section deals with both methods of handling errors.”. . 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. All rights reserved. transmitted. Send the reject stream to a *. 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 17.5 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.rej file. 3. Reject link exists. and should be reviewed by the Data Steward. No part of this publication may be reproduced. The appropriate file type.Rejects are tracked by count only. The Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job.rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504.7. The project phase (jpENVIRON) and a underscore “_”. and 5. For example.5. The job name (jpJOB_NAME) and a underscore “_”. The project name (jpPROJECT_NAME) and a underscore “_”.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. The job serial number (jpJOBSERIALNO) and a period “. stored in a retrieval system.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. 3. 2. 4. and gathered using a Funnel stage that feeds a Sequential File stage. or translated into any language in any form by any means without the written permission of IBM.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. 3.7. Only records that match the given table definition and database constraints are written. 2 Rejects should not exist but should not stop the job. transcribed. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. one of “rej” or “err”.

No part of this publication may be reproduced. the following stages process the only errors produced by a job: The Column Export stage maps the unique columns to the single standard column. Figure 13: Error Processing Components Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. 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. . All rights reserved.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. transmitted. The Transformer stage adds the required key columns. transcribed.

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

OR with columns contain illegal values for some operation performed on said columns. All rights reserved. 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. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key. 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. Connect the reject port to a Transformer stage where columns are set to specific values.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. transmitted. Connect the output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams. .7. In either case. The following table identifies the tagging method to be used for the previously cited operators. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream. transcribed.Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3.5. 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. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values. Switch Transformer In this example. or translated into any language in any form by any means without the written permission of IBM. 2006 50 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 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. Connect the reject port to a Transformer stage where columns are set to specific values. No part of this publication may be reproduced.

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

2 Copy Stage For complex data flows. No part of this publication may be reproduced.Dropping Columns July 17.8. 3. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run. For this reason. . it is important to minimize the number of transformers. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime.4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component.Job Design placeholder between stages (unless the Force option =true. • The Copy stage should be used instead of a Transformer for simple operations including: .BASIC Routines . in sort order. transcribed. Since the Copy stage does not require an output link.BASIC Transformers . 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.8. This is discussed in more detail in Parallel Framework Standard Practices: Administration. and Production Automation. stored in a retrieval system. All rights reserved. 3. without overhead for format or data type conversions.8. and to use other stages (such as Copy) when derivations are not needed. Management. it is best to develop a job iteratively using the Copy stage as a “placeholder”. transmitted. it can be used at the end of a data flow o For simple jobs with only two stages. Data Sets offer parallel I/O on read and write operations. and necessary.Renaming Columns .Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: . Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs. NOTE: Because parallel Data Sets are platform and configuration-specific. Used in this manner. or translated into any language in any form by any means without the written permission of IBM. and in Enterprise Edition native format. 3. Enterprise Edition will optimize this out at runtime) .3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs. o Unless the Force property is set to “True”. the Copy stage should be used as a placeholder so that new stages can be inserted easily should future requirements change. always write to parallel Data Sets.Server shared containers Note that BASIC Routines are still appropriate. they should not be used for long-term archive of source data.

1. Consider. Even then its use should be restricted and the routines should be converted as soon as possible. . and character string trimming.2. • Because the parallel Transformer is compiled.Complex reusable logic cannot be implemented using the Transformer or .Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4. the derivation expression: If A=0. transcribed. 3. stored in a retrieval system.6.Existing Transformers do not meet performance requirements As always. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible.2: Default and Explicit Type Conversions] Note that rename. if possible. 2006 53 of 179 © 2006 IBM Information Integration Solutions. and default type conversion can also be performed by the output mapping tab of any stage. implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. For example. Parallel Framework Red Book: Data Flow Job Design July 17.5 BuildOp stages BuildOps should only be used when: . or translated into any language in any form by any means without the written permission of IBM. drop (if Runtime Column Propagation is disabled).5. See Section 8. Instead. The only time that Filter or Switch should be used is when the selection clauses need to be parameterized at runtime. it is faster than the interpreted Filter and Switch stages. • NEVER use the “BASIC Transformer” stage in large-volume job flows. • The Modify stage can be used for non-default type conversions. transmitted. user-defined functions and routines can expand parallel Transformer capabilities. All rights reserved.3 Then B=”X” If A=4. No part of this publication may be reproduced.8. performance should this should be tested in isolation to identify specific cause of bottlenecks.1.2: Modify Stage.7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B. null handling. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines.

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. LongVarBinary. with resolution of microseconds (Specify microseconds Extended option) Single field containing both date and time value with resolution to seconds. VarBinary Unknown. Single field containing both date and time value with resolution to microseconds. Char. day. 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. LongNVarChar Char. LongVarChar. with resolution to seconds Time of day. uint16 int32. . Real Double TinyInt SmallInt Integer BigInt1 Binary. NVarChar. VarChar Time Time Timestamp Timestamp date decimal sfloat dfloat int8. (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. stored in a retrieval system. 2006 54 of 179 © 2006 IBM Information Integration Solutions. 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. and year Packed decimal. uint8 int16. VarChar NChar. All rights reserved. compatible with IBM packed decimal format.Information Integration Solutions Center of Excellence 4 DataStage Data Types The DataStage Designer and Manager represent column data types using SQL notation. transcribed. Numeric Float. Parallel Framework Red Book: Data Flow Job Design July 17. Bit. 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. transmitted. No part of this publication may be reproduced. LongVarChar. Each SQL data type maps to an underlying data type in the Enterprise Edition engine. unit32 int64. or translated into any language in any form by any means without the written permission of IBM. The following table summarizes the underlying data types of DataStage Enterprise Edition: SQL Type Internal Type Size Description Date Decimal.

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

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

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

for example the Sequential File stage. a fatal error occurs. the remaining columns pass through at runtime. 2006 58 of 179 © 2006 IBM Information Integration Solutions. allow their runtime schema to be parameterized further extending re-use through RCP. rather than using a large number of jobs with hard-coded table definitions to perform the same tasks. it must be enabled at the project level through the Administrator client. . 4. Furthermore. 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.3 Runtime Column Propagation Runtime column propagation (“RCP”) allows job designs to accommodate additional columns beyond those defined by the job developer. Some stages. transmitted. RCP facilitates re-use through parallel shared containers. All rights reserved. transcribed. Using RCP judiciously in a job design facilitates re-usable job designs based on input metadata.Information Integration Solutions Center of Excellence If the source value is null. as long as each stage in the shared container has RCP enabled on their stage Output properties. Using RCP. stored in a retrieval system. Before a DataStage developer can use RCP. Parallel Framework Red Book: Data Flow Job Design July 17. only the columns explicitly referenced within the shared container logic need to be defined.

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

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

although it can be explicitly set or cleared in the “Advanced” stage properties of a given stage. as shown in on the right: Figure 22: Preserve Partitioning option The Preserve Partitioning flag is part of the Data Set structure. 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. transmitted.1 Same Partitioning Same partitioning in fact performs no partitioning to the input Data Set. All rights reserved. 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. In these instances. Each partition receives the entire 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 some stages (Sort. and is appropriate when trying to preserve the grouping of a previous operation (for example. if the Preserve Partitioning flag was set. There are some cases when the input stage requirements prevent partitioning from being preserved. a parallel Sort). for example).1. stored in a retrieval system. No part of this publication may be reproduced. a warning will be placed in the Director log indicating that Enterprise Edition was unable to preserve partitioning for a specified stage.2.2 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. in the case of a cluster or Grid. it retains the partitioning from the output of the upstream stage. between servers). transcribed. Distributes rows evenly across partitions. Instead. Distributes rows evenly across partitions in a random partition assignment.1. and its state is stored in persistent Data Sets. Because Same does not redistribute existing partitions. in a round robin partition assignment. 2006 61 of 179 © 2006 IBM Information Integration Solutions. as illustrated on the right: Same partitioning doesn’t move data between partitions (or. 5. . the degree of parallelism remains unchanged: Parallel Framework Red Book: Data Flow Job Design July 17. 5.

as the complete Data Set must be distributed across the network to each node.3 Random Partitioning Like Round Robin. While in theory Random partitioning is not subject to regular data patterns that might exist in the source data. On clustered and Grid implementations. Random partitioning evenly distributes rows across partitions. 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. . the order that rows are assigned to a particular partition will differ between job runs. or translated into any language in any form by any means without the written permission of IBM.1. but using a random assignment. . Random partitioning has a slightly higher overhead than Round Robin partitioning. 3 2 1 0 . .2. regardless of the current $APT_CONFIG_FILE 5. …8 7 6 5 4 3 2 1 0 ENTIRE . it is rarely used in real-world data flows. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). . Entire partitioning may have a performance impact.1. similar to dealing cards: Round robin partitioning has a fairly low overhead. 3 2 1 0 .2. Since optimal parallel processing occurs when all partitions have the same workload. the downstream stage runs with the degree of parallelism used to create the Data Set.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially. 2006 62 of 179 © 2006 IBM Information Integration Solutions. Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning.2. As a result. Since the random partition number must be calculated.2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5. No part of this publication may be reproduced. stored in a retrieval system. 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17.4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition. All rights reserved. transmitted.1. transcribed.

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. For DB2 Enterprise Server Edition with DPF (DB2/UDB) only – matches the internal partitioning of the specified source or target table. Remove Duplicates) or stage requirements (for example. or translated into any language in any form by any means without the written permission of IBM. 2006 63 of 179 © 2006 IBM Information Integration Solutions. Join) require processing on groups of related records. Keyed partitioning is used when business rules (for example. No part of this publication may be reproduced.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.1. then the resulting partitions will be of relatively equal size.1. ensuring that records with the same values in those key column(s) are assigned to the same partition. stored in a retrieval system. All rights reserved.Information Integration Solutions Center of Excellence 5. As an example of hashing. and there are a large number of unique values.3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns. 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. transcribed.3. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. transmitted. If the source data values are evenly distributed within these key column(s). 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. 5. .

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

No part of this publication may be reproduced. The remainder is used to assign the value to a given partition: partition = MOD (key_value / number of partitions) Like hash. Also note that only the unique combination of key column values appear in the same partition when used for hash partitioning. When using hash partitioning on a composite key (more than one key column). stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. 5. transmitted. producing partition skew. there are more instances of “Ford” than “Dodge”. 2006 65 of 179 © 2006 IBM Information Integration Solutions.2 Modulus Partitioning Modulus partitioning uses a simplified algorithm for assigning related records based on a single integer key column. transcribed. the partition size of modulus partitioning will be equally distributed as long as the data values in the key column are equally distributed. .3. Parallel Framework Red Book: Data Flow Job Design July 17.1. individual key column values have no significance for partition assignment. All rights reserved. It performs a modulus operation on the data value using the number of partitions as the divisor. 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 in this example the number of unique values will limit the degree of parallelism.Information Integration Solutions Center of Excellence In this case. the key column combination of LName and FName yields improved data distribution and a greater degree of parallelism. regardless of the actual number of nodes in the parallel configuration file. Using the same source Data Set. which would impact performance.

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

4. The partition number is appended to the stage name for messages written to the Director log. 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.2Understanding the Parallel Job Score.2 Monitoring Partitions At runtime. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. and select the “Show Instances” option. transcribed. All rights reserved. Specific details on interpreting the parallel job score can be found in 12. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name). or translated into any language in any form by any means without the written permission of IBM. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True. right-click anywhere in the window. databases) c) and. in some cases. as illustrated below. and each stage is processing an equal number (12. 2006 67 of 179 © 2006 IBM Information Integration Solutions. transmitted. In this instance. 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. starting at zero. 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. . This is very useful in determining the distribution across parallel partitions (skew). No part of this publication may be reproduced.500) of rows for an optimal balanced workload. if specified. Partitions are assigned numbers.Information Integration Solutions Center of Excellence 5.

minimizing overall run time. 5. All rights reserved. Director. . This ensures that the processing workload is evenly balanced.3 Partition Methodology Given the numerous options for keyless and keyed partitioning. 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. No part of this publication may be reproduced. stored in a retrieval system. the orchadmin command line utility on the DataStage server can also be used to examine a given parallel Data Set. as shown below: Figure 27: Data Set Management Tool In a non-graphical way. 2006 68 of 179 © 2006 IBM Information Integration Solutions. while minimizing overhead.Information Integration Solutions Center of Excellence Figure 26: Output of APT_RECORD_COUNTS in Director log Finally. the “Data Set Management” tool (available in the Tools menu of Designer. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. or Manager) can be used to identify the degree of parallelism and number of rows per partition for an existing persistent Data Set. transmitted. or translated into any language in any form by any means without the written permission of IBM.

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

5.4.4.2Sorting and Hashing Advanced Example. we can move the Hash partition and Sort before the Copy stage.4. All rights reserved. more advanced partitioning and sorting examples are given in 12. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row. stored in a retrieval system. the partitioning and sorting of this scenario can be optimized.Information Integration Solutions Center of Excellence 5. . a Transformer is used to extract data from a single header row of an input file. transmitted. on closer inspection. a new output column is defined on the header and detail links using a single constant value derivation.4 Partitioning Examples In this section. or through Auto partitioning): 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. Using a “standard” solution. we’ll apply the partitioning methodology defined earlier to several example job flows. 2006 70 of 179 © 2006 IBM Information Integration Solutions.2 Partitioning Example 2 – Use of Entire Partitioning In this example. To add aggregate columns to every detail row. The standard solution would be to Hash partition (and Sort) the inputs to the Join and Aggregator stages as shown below: Figure 28: “Standard” Partitioning assignment However. as shown below: Figure 29: Optimized Partitioning assignment This example will be revisited in the Sorting discussion because there is one final step necessary to optimize the sorting in this example. Additional. transcribed. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly. Because the Join and Aggregator use the same partition keys and sort order. and apply Same partitioning to the downstream links.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. The output of the Aggregator is then sent to the second input of the Join. 5. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator. No part of this publication may be reproduced. Within the Transformer.

No part of this publication may be reproduced. 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). 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. Although functionally correct. depending on the business requirements. Remembering that the degree of parallel operation is limited by the number of distinct values.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions . there is no need to pre-sort the input to the Join. so we will revisit this in the Sorting discussion. the link order of the Inner Join is significant. For this reason. An optimized solution would be to alter the partitioning for the input links to the Join stage: . To optimize partitioning. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. All rights reserved. the above solution has one serious limitation. In order to process a large number of detail records. transmitted. . 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. 2006 71 of 179 © 2006 IBM Information Integration Solutions. resulting in sequential processing. consider that the single header row is really a form of reference data.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.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. transcribed. the single value join column will assign all rows to a single partition.

All rights reserved. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence If defined in reverse of this order. there is a specialized example where the Round Robin collector may be appropriate. However. 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. there is one further detail in this example.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. or translated into any language in any form by any means without the written permission of IBM. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory. 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 collector method is defined in the stage Input/Partitioning properties for any stage running sequentially. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition. 5. stored in a retrieval system. 5. the order of rows in an Auto collector is undefined. 2006 72 of 179 © 2006 IBM Information Integration Solutions.5. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5.3: Minimizing Runtime Processes and Resource Requirements). Like partitioning methods.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially. and may vary between job runs on the same Data Set.5. 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. (See Section 12. For advanced users. transmitted. For this reason. transcribed. the detail rows in the Left input will buffer to disk to prevent a deadlock. Auto is the default collector method. .

the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. or translated into any language in any form by any means without the written permission of IBM. transcribed.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. Parallel Framework Red Book: Data Flow Job Design July 17.3 Ordered Collector An Ordered collector reads all rows from the first partition. in the same order. use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. transmitted. use Sort Merge collector to produce a single.5. and these should be the same columns. All rights reserved. In this scenario. 5.5. .4 Sort Merge Collector If the input Data Set is sorted in parallel. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. 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. 2006 73 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. 5. the Sort Merge collector will generate a sequential stream of rows in globally sorted order. 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). stored in a retrieval system. an Ordered collector will generate a sequential stream in sort order. This is because Round Robin collector reads from partitions using the same partition order that a Round Robin partitioner assigns rows to parallel partitions. as long as the Data Set has not been repartitioned or reduced. The Sort Merge collector requires one or more key column(s) to be defined. through aggregation).6 Collecting Methodology Given the options for collecting data into a sequential stream. then reads all rows from the next partition until all rows in the Data Set have been collected. 5. Row order is undefined for non-key columns. as used to sort the input Data Set in parallel.

Partitioning: is used to gather related records. assigning rows with the same key column values to the same partition . the process of sorting data uses one primary key column and. 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. 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. Other stages (for example. This sort can be done in parallel. For example. transcribed. 6. Each column is specified with an ascending or descending sort order. 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. stored in a retrieval system. optionally. The order of key columns determines the sequence and groupings in the result set. No part of this publication may be reproduced. ordered result set. sorting on primary key LName (ascending). 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. All rights reserved. a sort is required after repartitioning. transmitted. we would see the following results: Parallel Framework Red Book: Data Flow Job Design July 17. Change Apply. 2006 74 of 179 © 2006 IBM Information Integration Solutions. as illustrated in the following example. Join. sort order is not maintained. based on one or more key column(s) NOTE: By definition. Instead. To restore row order and groupings. Change Capture. or translated into any language in any form by any means without the written permission of IBM.1 Partition and Sort Keys Using the parallel Sort within DataStage Enterprise Edition: .Information Integration Solutions Center of Excellence 6 Sorting Traditionally. in most cases there is no need to globally sort data to produce a single sequence of rows. Merge) require pre-sorted groups of related records. Sort Aggregator.Sorting: is used to establish group order within each partition. Given a 4-node configuration file. . when data is re-partitioned. In the following example. This is the method the SQL databases use for an ORDER BY clause. one or more secondary key column(s) to generate a sequential.

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

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

. Always specify “DataStage” Sort Utility.Information Integration Solutions Center of Excellence - Position of nulls in the result set (for nullable columns) Figure 36: Specifying Link Sort options 6. the Sort Key Mode is most frequently used. 2006 77 of 179 © 2006 IBM Information Integration Solutions. which is significantly faster than a “UNIX” sort. transcribed. and can often be performed in memory (depending on the size of the new subsort groups). Specifically. or translated into any language in any form by any means without the written permission of IBM. For this reason. Parallel Framework Red Book: Data Flow Job Design July 17. It is used when resorting a sub-grouping of a previously sorted input Data Set. 6. This requires some additional overhead in the sort algorithm. All rights reserved.3. This “subsort” uses significantly less disk space and CPU resource. the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys.5 Sub-Sorts Within the standalone Sort stage. No part of this publication may be reproduced. 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. instead of performing a complete Sort. stored in a retrieval system.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. NOTE: The Sort Utility option is an artifact of previous releases.4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. - 6. transmitted. 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. disable Stable sort unless needed.

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

. 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. All rights reserved. By default. or translated into any language in any form by any means without the written permission of IBM. 2006 79 of 179 © 2006 IBM Information Integration Solutions. Create Key Change Column. There are two ways to prevent Enterprise Edition from inserting an un-necessary sort: a) Insert an upstream Sort stage on each link. and disk resources. rows in the input Data Set are read into a memory buffer on each partition. On a global basis. transmitted. 6. This will verify sort order but not actually perform a sort. sequential ordered result set use a parallel Sort and a Sort Merge collector 6. each sort uses 20MB of memory per partition for its memory buffer. No part of this publication may be reproduced. This value can be changed for each standalone Sort stage using the “Restrict Memory Usage” option (the minimum is 1MB/partition). the environment variable APT_TSORT_STRESS_BLOCKSIZE can be Parallel Framework Red Book: Data Flow Job Design July 17. memory.8 Tuning Sort Sort is a particularly expensive task within DataStage Enterprise Edition which requires CPU. inserted sorts can be a significant performance impact if they are not necessary. To perform a sort. If the sort operation can be performed in memory (as is often the case with a subsort) then no disk I/O is performed. 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. Revisiting the partitioning examples in Section 5. stored in a retrieval system.Information Integration Solutions Center of Excellence While ensuring correct results. Create Cluster Key Change Column.7 Sort Methodology Using the rules and behavior outlined in the previous section. the environment variable $APT_SORT_INSERTION_CHECK_ONLY should be set to prevent Enterprise Edition from inserting un-necessary sorts before the Join stage. 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. transcribed.4: Partitioning Examples. aborting the job if data is not in the required sort order.

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

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

transmitted. Unlike the Complex Flat File stage.3. define a single large string column for the non-parallel Sequential File read.file may be either fixed or variable-width Read Method: Specific Files.3.file may only be fixed-width Note that when reading in parallel. Read From Multiple Nodes option is set to Yes useful for cluster and Grid configurations . .3. stored in a retrieval system.Information Integration Solutions Center of Excellence 7. and then pass this to a Column Import stage to parse the file in parallel.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. input row order is not maintained across readers. or translated into any language in any form by any means without the written permission of IBM.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. only one file specified. Readers Per Node option greater than 1 useful for SMP configurations . 7. The FileSet stage will write in parallel. the degree of parallelism of the write will correspond to the number of file names specified. 2006 82 of 179 © 2006 IBM Information Integration Solutions.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. No part of this publication may be reproduced. using the format options specified in the FileSet stage. performance can still be improved by separating the file I/O from the column parsing operation. 7.3 Separating I/O from Column Import If the Sequential File input cannot be read in parallel. All rights reserved. transcribed. in parallel. This will create a single header file (in text format) and corresponding data files. set environment variable $APT_IMPORT_PATTERN_USES_FILESET Read Method: Specific Files. 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. As shown in the following Job fragment. 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. 7. Parallel Framework Red Book: Data Flow Job Design July 17. In these instances. A better option for writing to a set of Sequential Files in parallel is to use the FileSet stage. more than one file specified each file specified within a single Sequential File stage must be of the same format Read Method: File Pattern.

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

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

but you can configure the stage to execute sequentially if it is only reading one file with a single reader. No part of this publication may be reproduced. . transcribed. Increasing this size may improve performance. Complex Flat File source stages execute in parallel mode when they are used to read multiple files. the stage allows you to write data to one or more complex flat files.3.10 Tuning Sequential File Performance On heavily-loaded file servers or some RAID/SAN array configurations. or OCCURS clauses. these extra characters will be silently truncated. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. REDEFINES. the environment variables $APT_IMPORT_BUFFER_SIZE and $APT_EXPORT_BUFFER_SIZE can be used to improve I/O performance. 2006 85 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Varchar length. A complex flat file may contain one or more GROUPs. 7. transmitted. the stage allows you to read data from one or more complex flat files. It does not write to MVS datasets.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. NOTE: The Complex Flat File stage cannot read from sources with OCCURS DEPENDING ON clauses. 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. 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. All rights reserved. When used as a source. These settings specify the size of the read (import) and write (export) buffer size in Kbytes. 7.) When used as a target. Finally. including MVS datasets with QSAM and VSAM files. (This is an error in the DataStage documentation. with a default of 128 (128K). in some disk array configurations.

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

Always include reject links in a parallel Transformer. .col) Then “” Else link. if you intend to use a nullable column within a Transformer derivation or output link constraint. transmitted.8. 2006 87 of 179 © 2006 IBM Information Integration Solutions. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17. See Section 3. No part of this publication may be reproduced. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link. 8. To create a Transformer reject link in Designer. undefined.1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. the Transformer will reject the row instead of sending it to the output link(s).col Note that if an incoming column is only used in an output column mapping. For this reason. stored in a retrieval system. 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. Even if the target column in an output derivation allows nullable results. 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. transcribed.Information Integration Solutions Center of Excellence 8 Transformation Languages 8.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints.1. entries are placed in the Director job log. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. When rows are rejected by a Transformer. 8. Because the DS/EE Transformer runs in parallel.1. the Transformer will allow this row to be sent to the output link without being rejected. and to use other stages (such as Copy) when derivations are not needed.4: Parallel Transformer stages for guidelines on Transformer stage usage. All rights reserved. For example. it is important to minimize the number of transformers. For this reason. by definition.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer. or translated into any language in any form by any means without the written permission of IBM. This makes it easy to identify reject conditions (by row counts).

Examples: 1. Since the Transformer will abort the entire job flow immediately. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7. there are times when internal decimal variables need to be generated in order to perform the evaluation.4 -> -1. -1. Examples: 1. 8. the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range.1. Examples: 1.4 -> -2 round_inf Rounds or truncates towards nearest representable value. TrimLeadingTrailing(string) works only if string is a VarChar field. apply the abort rule to this output link. 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.1. it is possible that valid rows will not have been flushed from Sequential File (export) buffers.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition. By default. Therefore. 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. the incoming column must be type VarChar before it is evaluated in the Transformer. not the target. 1). No part of this publication may be reproduced. When generating a sequence of numbers in parallel. . the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example. Create a new output link that will handle rows that match the abort rule. up to a maximum precision of 255 and scale of 125.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. 8. it is important to make sure the type conversion is done before a row reaches the Transformer. the PadString function uses the length of the source type. For example.4 -> 2. Thus. All rights reserved.1. -1. Within the link constraints dialog box.3. 1.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment. these internal decimal variables will have a precision and scale of 38 and 10.4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage. By default. 2006 88 of 179 © 2006 IBM Information Integration Solutions. If more precision is required.4 -> 1. For example. internal decimal results are rounded to the nearest applicable value. -1.5: Sequential File (Export) Buffering).5-> 2.6 ->1. transmitted. or committed to database tables. -1. 8. stored in a retrieval system. or performing parallel derivations.6 -> -1 floor Rounds towards negative infinity.

. Examples: 1. The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value.1. the substring is evaluated just once for every input row. as they would be evaluated once for every output column that uses them. 2006 89 of 179 © 2006 IBM Information Integration Solutions. the evaluation of the substring of DSLINK1. then the following test may appear in a number of output columns derivations: IF (DSLINK1.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations. In this case. All rights reserved. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. stored in a retrieval system.5. In this case. No part of this publication may be reproduced. round or truncate to the scale size. From this sequence. By doing this.col[1. For example.col[1. it is useful to understand what items get evaluated and when.3] is evaluated for each column that uses it. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example.5. Similarly. -1. it can be seen that there are certain constructs that would be inefficient to include in output column derivations. the stage variable definition would be: Parallel Framework Red Book: Data Flow Job Design July 17..56 ->-1.3] = “001”) THEN . This can be made more efficient by moving the substring calculation into a stage variable. transcribed.56 -> 1. the output links are also evaluated in the order in which they are displayed.. unless the derivation is empty For each output link: Evaluate the link constraint. suppose multiple columns in output links want to use the same substring of an input column.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign. 8. transmitted. or translated into any language in any form by any means without the written permission of IBM. 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.

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

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

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

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

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

An exception to this is financial calculations which should be done in decimal to preserve appropriate precision. which maintains the results of each keycolumn value/aggregation pair in memory.3. set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. transmitted.3. the output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double). 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 default Aggregation Method is set to Hash. 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. To perform a total aggregation. Parallel Framework Red Book: Data Flow Job Design July 17. To aggregate in decimal precision. All rights reserved. No part of this publication may be reproduced. The Sort Aggregation Method should be used when the number of key values is unknown or very large. Because each key value/aggregation requires approximately 2K of memory.Information Integration Solutions Center of Excellence 9. transcribed. or translated into any language in any form by any means without the written permission of IBM. .1 Aggregation Method By default. Unlike the Hash Aggregator. 9.3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings.3.3 The Aggregator Stage 9.2 Aggregation Data Type By default. 9. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property. aggregate on generated key column) there is no need to sort or hash-partition the input data with only one key column value aggregate Sequentially on the generated column - - Note that in this example use two Aggregators are used to prevent the sequential aggregation from disrupting upstream processing. the Hash Aggregator should only be used when the number of distinct key values is small and finite. stored in a retrieval system. but only maintains the calculations for the current group in memory. the Sort Aggregator requires presorted data. 2006 95 of 179 © 2006 IBM Information Integration Solutions.

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

transcribed.1. 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. Parallel Framework Red Book: Data Flow Job Design July 17. Because plug-in stage types cannot read in parallel.Information Integration Solutions Center of Excellence 10. attributes). and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied. operating system. not name. Unlike the database-specific parallel stages. plug-in database stages match columns by order.1. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform. On UNIX. the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server.1. the DataDirect ODBC Driver Manager is also included.2 ODBC Enterprise stage In general. Because there are exceptions to this rule (especially with Teradata). types. 2006 97 of 179 © 2006 IBM Information Integration Solutions. the native parallel stages often have more stringent connectivity requirements than plug-in stages. 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). stored in a retrieval system. so Table Definitions must match the order of columns in a query. and version.1 Native Parallel database stages In general. For example. Furthermore. This allows Enterprise Edition to match return columns by name. they should only be used when it is not possible to use a native parallel stage. All rights reserved. for maximum parallel performance. specific guidelines of when to use various stage types are provided in the database-specific topics in this section. 10.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. 10. transmitted. not position in the stage Table Definitions.1. the ODBC Enterprise stage cannot interface with database-specific parallel load technologies. No part of this publication may be reproduced. care must be taken to assign the correct data types in the job design. native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform. Native parallel stages always pre-query the database for actual runtime metadata (column names. Because of their tight integration with database technologies. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. or translated into any language in any form by any means without the written permission of IBM. scalability. and cannot span multiple servers in a clustered or Grid configuration. .1. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect. From a design perspective. However.1.

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. However. and selecting “Import from Database Table” option in the wizard as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. Informix Enterprise or Oracle Enterprise stages.2 Database Metadata 10. .2. transcribed.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 Runtime metadata At runtime. This utility is available as a server command line utility and within Designer and Manager using “Import Orchestrate Schema Definitions”. For each native parallel database stage: . data types.Information Integration Solutions Center of Excellence 10. This allows the database stages to match return values by column name instead of position. transmitted.1. or translated into any language in any form by any means without the written permission of IBM.1. 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.both DS/EE and relational databases support null values. Database-specific data type mapping tables are included in the following sections. the DS/EE native parallel database stages always “pre-query” the database source or target to determine the actual metadata (column names. 2006 98 of 179 © 2006 IBM Information Integration Solutions.rows of the database result set correspond to records of a DS/EE Data Set . regardless of the table definitions assigned by the DataStage developer. 10. stored in a retrieval system.2. All rights reserved. nullability) and partitioning scheme (in some cases) of the source or target table.columns of the database row correspond to columns of a DS/EE record . use orchdbutil to import metadata to avoid type conversion issues.1.2 Metadata Import When using the native parallel DB2 Enterprise.

2006 99 of 179 © 2006 IBM Information Integration Solutions.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. 10.2. . the DataStage Designer will automatically populate the select list based on the stage’s output column definition.3 Optimizing Select Lists For best performance and optimal memory usage. you will see a window with the result columns and data. 10. The alias name(s) should then be added to the Table Definition within DataStage. transmitted. similar to the illustration on the right: Parallel Framework Red Book: Data Flow Job Design July 17. it will be easier to use the corresponding orchdbutil command-line utility from the DataStage server machine. As a command.1.1. Test the connection using View Data button. For “Auto-Generated” SQL. orchdbutil can be scripted to automate the process of importing a large number of tables. 10. If the connection is successful. it is best to explicitly specify column names on all source database stages. always specify the “Select List” subproperty. stored in a retrieval system. All rights reserved. instead of using an unqualified “Table” or SQL “SELECT *” read. the following SQL assigns the alias Total to the calculated column: SELECT store_name. or translated into any language in any form by any means without the written permission of IBM. For example. there may be cases where user-defined functions or logic need to be executed on the database server. it is important to use SQL aliases to explicitly name the calculated columns so that they can be referenced within the DataStage job. For “Table” read method. transcribed. No part of this publication may be reproduced.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. 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.3 Defining Metadata for Database Functions When using database functions within a SQL SELECT list in a Read or Lookup.1. When importing a large number of tables. 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.

or translated into any language in any form by any means without the written permission of IBM. 10. the results should be landed to a parallel Data Set. All rights reserved. it is not a good idea to let DataStage generate target tables unless they are used for temporary storage. the OPEN command can be used to create a target table.6 Database OPEN and CLOSE Commands The native parallel database stages provide options for specifying OPEN and CLOSE commands. including databasespecific options (tablespace. For example. constraints. As a further optimization. the OPEN command could be used to create a temporary table. and the CLOSE command could be used to select all rows from the temporary table and insert into a final target table. transmitted. Subsequent job(s) should read this Data Set and populate the target table using the appropriate database stage and write method. There are limited capabilities to specify Create table options in the stage. As another example. a Lookup stage (or Join stage. an error message may appear. transcribed.5 Designing for Restart To enable restart of high-volume jobs. . Parallel Framework Red Book: Data Flow Job Design July 17.1. depending on data volume) can be used to identify existing rows before they are inserted into the target table. After transformation. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 43: Sample View Data Output If the connection fails. and doing so may violate data-management (DBA) policies. stored in a retrieval system.1. and you will be prompted to view additional detail. Figure 44: View Additional Error Detail 10. In general. 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. etc) not possible with the “Create” option. 2006 100 of 179 © 2006 IBM Information Integration Solutions. OPEN and CLOSE are not offered by plugin database stages. logging. 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. it is important to separate the transformation process from the database write (Load or Upsert) operation.

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

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

2.2. For this reason. If you use the stage in this way. (Before DB2 v8. consult the DataStage documentation (DataStage Parallel Job Developers Guide. 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.1 DB2/UDB Enterprise stage Enterprise Edition provides native parallel read. This will only work when the authentication mode of the database on the remote instance is set to “client authentication”. the “client authentication” configuration of a remote instance is not recommended. 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. transcribed. stored in a retrieval system. Check with IBM IIS Support for availability.) Furthermore.1. transmitted. . and DataStage Plug-In guides) 10. or translated into any language in any form by any means without the written permission of IBM. The DB2/UDB Enterprise stage requires DB2 Enterprise Server Edition on UNIX with Data Partitioning Facility (DPF) option. this was also called “DB2 EEE”. No part of this publication may be reproduced. 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. 2006 103 of 179 © 2006 IBM Information Integration Solutions. 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. and load capabilities to parallel DB2 databases on UNIX using the native parallel DB2/UDB Enterprise stage.1 DB2 Stage Types DataStage Enterprise Edition provides access to DB2 databases using one of 5 stages. 3 A patched version of the ODBC Enterprise stage allowing parallel read is available from IBM IIS Support for some platforms. lookup. Parallel Framework Red Book: Data Flow Job Design July 17.2 DB2 Guidelines 10. the DB2 hardware/UNIX/software platform must match the hardware/software platform of the DataStage ETL server. upsert.Information Integration Solutions Center of Excellence 10. All rights reserved.

2.2 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. By facilitating flexible connectivity to multiple types of remote DB2 database servers. databases on UNIX platforms that differ from the platform of the DataStage ETL server. including direct communication with each DB2 database node. All rights reserved.1. transcribed. using the same data partitioning as the referenced DB2 tables. plug-in stages cannot read from DB2 in parallel. 2006 104 of 179 © 2006 IBM Information Integration Solutions. 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).2. the ability to write in parallel may be limited by the table and index configuration set by the D2 database administrator. 10. transmitted.writing to DB2 in parallel (where appropriate). 10. the use of DataStage plug-in stages will limit overall performance and scalability. . parallel component the DB2/UDB Enterprise stage is designed for maximum performance and scalability. These stages also provide connectivity to non-UNIX DB2 databases. No part of this publication may be reproduced. These goals are achieved through tight integration with the DB2 RDBMS. Add the following properties: Parallel Framework Red Book: Data Flow Job Design July 17. 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). it may be possible to write to a DB2 target in parallel.Information Integration Solutions Center of Excellence As a native. 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. and reading from . Because each plug-in invocation will open a separate connection to the same target DB2 database table. or translated into any language in any form by any means without the written permission of IBM. Sparse Lookup is not supported through the DB2/API stage. Furthermore.2 Connecting to DB2 with the DB2/UDB Enterprise Stage Create a Parallel job and add a DB2/UDB Enterprise stage. Using the DB2/UDB API stage or the Dynamic RDBMS stage. when used as data sources. “DB2 EE”).

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

Since the APT_DB2INSTANCE_HOME is not set. DB2 Load.5 DB2/API stage Column Names When using the DB2/API. this will create problems.cfg. set the environment variable $DS_ENABLE_RESERVED_CHAR_CONVERT if your DB2 database uses the reserved characters # or $ in column names. 2006 106 of 179 © 2006 IBM Information Integration Solutions.2. .the name must start with a letter or underscore character .4 DB2/UDB Enterprise stage Column Names At runtime. No part of this publication may be reproduced. 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). DS will default to this directory to find the configuration file for the remote instance. • In order for DataStage to locate the db2nodes.the name is case insensitive When there is an incompatibility. the string “APT__column#” (two underscores) is added to beginning of the column name. where column# is the number of the column. Dynamic RDBMS) may be used to access data in other instances. DB2 Load.if the DB2 column name contains a character that is not alphanumeric or an underscore. and Dynamic RDBMS plug-in stages. In the users UNIX home directory. Observe the following guidelines when $DS_ENABLE_RESERVED_CHAR_CONVERT is set: Parallel Framework Red Book: Data Flow Job Design July 17. which place no limit on the length of a column name. To connect to multiple DB2 instances. the character is replaced by two underscore characters 10. although this incurs the overhead of Sequential File stage (corresponding export/import operators) which does not run in parallel.the name can contain only alphanumeric and underscore characters . For example. Depending on platform configuration and I/O subsystem performance. the Enterprise Edition column will be named “APT_37dig” . if the data volumes are sufficiently small. 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. All rights reserved. transmitted. 10.2. landing intermediate results to a parallel Data Set. transcribed. separate jobs can communicate through named pipes. DB2 plug-in stages (DB2 API. but have the following restrictions: . Enterprise Edition converts the DB2 column name as follows: . we recommend using separate jobs with their respective DB2 environment variable settings.if the DB2 column name does not begin with a letter or underscore.Information Integration Solutions Center of Excellence only contain information for one instance. stored in a retrieval system. This converts these special characters into an internal representation that DataStage can understand. if the third DB2 column is named 7dig. or translated into any language in any form by any means without the written permission of IBM.cfg there. Or. create a sqllib subdirectory and place the remote instance’s db2nodes.

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

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

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

transmitted. No part of this publication may be reproduced. 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. An example of an in-memory Normal Lookup is shown in Figure 48. . transcribed.Information Integration Solutions Center of Excellence Figure 47: DB2 write on DataStage/USS On DataStage/USS Edition. 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. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 110 of 179 © 2006 IBM Information Integration Solutions.

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

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

and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. If the DATETIME starts with a year component. Informix Data Type CHAR(n) CHARACTER VARYING(n.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 .rows of the database result set correspond to records of a DS/EE Data Set . transmitted. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. as shown in the following table. 2006 113 of 179 © 2006 IBM Information Integration Solutions.3. .2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.Information Integration Solutions Center of Excellence 10.r) NVARCHAR(n. No part of this publication may be reproduced. stored in a retrieval system.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: .both DS/EE and Informix support null values. decimal[p. All rights reserved. time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month. the result is a date field. the result is a time field. The Informix Enterprise stage converts Informix data types to Enterprise Edition data types. the result is a timestamp field. DECIMAL[p.3 Informix Database Guidelines 10.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour.s] where p is the precision and s is the scale The maximum precision is 32. transcribed.columns of the database row correspond to columns of a DS/EE record . 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.3.r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date. or translated into any language in any form by any means without the written permission of IBM.

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

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. stored in a retrieval system. 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. . No part of this publication may be reproduced. transmitted. transcribed. 2006 115 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.

All rights reserved. since this capability is not provided by the ODBC API. . or translated into any language in any form by any means without the written permission of IBM.3 Reading ODBC Sources Unlike other native parallel database stages. No part of this publication may be reproduced. Depending on the target database. stored in a retrieval system. Parallel reads through ODBC match the degree of parallelism in the $APT_CONFIG_FILE. 4 On some platforms.Information Integration Solutions Center of Excellence 10. 2006 116 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. a patch may be available through IBM IIS Support to support parallel reads through ODBC. and the table configuration (row or page level lock mode if available). the ODBC Enterprise stage does not support parallel read4. it may be possible to write to a target database in parallel using the ODBC Enterprise stage.4. transcribed.

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 . the Oracle Enterprise stage reads sequentially from its source table or query.Information Integration Solutions Center of Excellence 10.rows of the database result set correspond to records of a DS/EE Data Set . Setting the partition table option to the specified table will enable parallel extracts from an Oracle source.names are translated exactly except when the Oracle source column name contains a character that DataStage does not support. stored in a retrieval system. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage. In that case.5.5. The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition.queries containing a GROUP BY clause that are also hash partitioned on the same field Parallel Framework Red Book: Data Flow Job Design July 17.2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.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. No part of this publication may be reproduced.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p. or translated into any language in any form by any means without the written permission of IBM. If you attempt to read a record larger than 32K. transcribed.5 Oracle Database Guidelines 10.1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. transmitted. two underscore characters replace the unsupported character . All rights reserved. 2006 117 of 179 © 2006 IBM Information Integration Solutions.columns of the database row correspond to columns of a DS/EE record .both DS/EE and Oracle support null values. Examples include: . Enterprise Edition will return an error and abort your job. which are not heavily typed. Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p. This is particularly important for Oracle databases.s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38.3 Reading from Oracle in Parallel By default. and will generate an error at runtime 10. .

transmitted.5. • Setting the environment variable $APT_ORACLE_LOAD_OPTIONS to “OPTIONS (DIRECT=TRUE. Parallel Framework Red Book: Data Flow Job Design July 17. .4 Oracle Load Options When writing to an Oracle table (using Write Method = Load). stored in a retrieval system. 2006 118 of 179 © 2006 IBM Information Integration Solutions. the key column(s) must be identified using the check boxes in the column grid. the Oracle load will be done sequentially. PARALLEL=FALSE) also allows loading of indexed tables without index maintenance. When using this method. 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. 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. rebuild). 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 this instance. transcribed. All rights reserved. In order to automatically generate the SQL required by the Upsert method.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.

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.

10. 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.password=password[. where pa$$ is the example password: \’pa$$\’ 10.7. RequestedSessions equals the maximum number of available sessions on the Teradata instance. and hence the number of UNIX processes and overall system resource requirements of the DataStage job. By default.7.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. Teradata Enterprise stage uses 32K buffers.SessionsPerPlayer=nn][. To specify a Teradata password that contains special characters. or translated into any language in any form by any means without the written permission of IBM. and will generate an error at runtime. transcribed. the password must be surrounded by an “escaped” single quote as shown. Indirectly.7. 10. No part of this publication may be reproduced. the DB Options property specifies the connection string and connection properties in the form: user=username.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. In that case SessionsPerPlayer should be increased.6 Specifying Teradata Passwords with Special Characters Teradata permits passwords with special characters and symbols.RequestedSessions=nn] where SesionsPerPlayer and RequestedSessions are optional connection parameters that are required when accessing large Teradata databases. 2006 122 of 179 © 2006 IBM Information Integration Solutions. The SessionsPerPlayer option determines the number of connections each DataStage EE player opens to Teradata. (Note that 64K buffers must be enabled at the Teradata server level). . Setting the SessionsPerPlayer too low on a large system can result in so many players that the job fails due to insufficient resources. transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. and/or RequestedSessions should be decreased.7 Teradata Enterprise Settings Within the Teradata Enterprise stage. this determines the number of DataStage players. Aggregates and most arithmetic operators are not allowed in the SELECT clause of a Teradata Enterprise stage. but this can be set to a value between 1 and the database vprocs. All rights reserved. By default. stored in a retrieval system.

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

If the job results are correct with a single-node configuration file. the job’s partitioning logic and parallel design concepts (especially within Transformer stages) should be examined. This environment variable should not normally need to be set. This can be useful when determining if the actual runtime schema matches the expected job design table definitions. lowering this number will limit the number of processes that are simultaneously created when a job is started. Prints detailed information in the job log for each operator when allocating additional heap memory. Normally. 2006 124 of 179 $APT_PM_PLAYER_TIMING 1 1 FORCE $APT_PM_PLAYER_MEMORY $APT_BUFFERING_POLICY Setting $APT_BUFFERING_POLICY=FORCE is not recommended for production job runs. transmitted. . stored in a retrieval system. 11. including CPU utilization and elapsed processing time. When trying to start very large jobs on heavilyloaded servers. No part of this publication may be reproduced. Using the job monitor performance statistics. $DS_PX_DEBUG 1 $APT_PM_STARTUP_CONCURRENCY 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. and incorrect with a multi-node configuration file. July 17.1 Warning on Single-Node Configuration Files Because the DS/EE configuration file can be changed at runtime. the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance. 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. Forces an internal buffer operator to be placed between every operator. Prints detailed information in the job log for each operator. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. Set this environment variable to capture copies of the job score.Information Integration Solutions Center of Excellence 11 Troubleshooting and Monitoring 11. All rights reserved. transcribed. This will ensure that the jobs have been designed with proper partitioning logic. and internal Enterprise Edition log messages in a directory corresponding to the job name. or translated into any language in any form by any means without the written permission of IBM.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. this can identify which part of a job flow is impacting overall performance. generated osh.

look at row counts. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. stored in a retrieval system. Set the environment variable $DS_PX_DEBUG to capture copies of the job score. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG .Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. and internal Enterprise Edition log messages in a directory corresponding to the job name. All rights reserved. it is possible for $OSH_PRINT_SCHEMAS to generate a log entry that is too large for DataStage Director to store or display. Use the Data Set Management tool (available in the Tools menu of DataStage Designer or DataStage Manager) to examine the schema. or promote a job from development into test or production. The default is 30 seconds. the log entry is sometimes preceded by a warning condition. .  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled). When a fatal error occurs. 11. transmitted.3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. and to manage source or target Parallel Data Sets. generated osh.  Enable the Job Monitoring Environment Variables detailed in Section 2.5. In some instances. tune. 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. To capture the full schema output in these cases. transcribed. These may indicate an underlying logic problem or unexpected data type conversion. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. The general process for debugging a job is:  Check the Director job log for warnings. But all warnings should be examined and understood. All fatal and warning messages should be addressed before attempting to debug.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide.    NOTE: For large jobs. July 17. it may not be possible to remove all warning messages generated by the EE engine. No part of this publication may be reproduced. This will place entries in the Director log with the actual runtime schema for every link using Enterprise Edition internal data types. or translated into any language in any form by any means without the written permission of IBM.

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

5 Interpreting the Parallel Job Score When attempting to understand an Enterprise Edition flow. that is. the first task is to examine the score dump which is generated when you set APT_DUMP_SCORE=TRUE in your environment.Information Integration Solutions Center of Excellence Figure 52: Generated OSH in Designer Job Properties 11. It shows three stages: Generator. transmitted. The Peek and Sort stages are combined. stored in a retrieval system. Also available is some information about where data may be buffered. A score dump includes a variety of information about a flow. what degree of parallelism each operator runs with. or translated into any language in any form by any means without the written permission of IBM. which operators. No part of this publication may be reproduced. 2006 127 of 179 © 2006 IBM Information Integration Solutions. The job runs 3 processes on 2 nodes. which has a hash partitioner that partitions on key field a. have been inserted by EE. ##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. The following score dump shows a flow with a single Data Set. including how composite operators and shared containers break down. if any. All stages in this flow are running on one node. Sort (tsort) and Peek. and exactly which nodes each operator runs on. transcribed. . they have been optimized into the same process. All rights reserved. where data is repartitioned and how it is repartitioned.

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.4. stored in a retrieval system. . are the computation-intensive stages shared evenly across all nodes? More details on interpreting the parallel job score can be found in 12.com[op1.p0] )} op1[2p] {(parallel APT_CombinedOperatorController: (tsort) (peek) )on nodes ( lemond.com[op1.torrent.p1] )} In a score dump. transcribed.p0] lemond. 2006 128 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence lemond.torrent. No part of this publication may be reproduced. All rights reserved. transmitted.2 Understanding the Parallel Job Score.torrent.com[op0. Parallel Framework Red Book: Data Flow Job Design July 17.

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

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

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

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

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

BuildOps.4. Cannot be less than 8192. DB2 Enterprise. Remove Duplicates. Default is 8192. with a valid value range for between 8192 and 1048576. Merge. cannot be greater than 1048576. transcribed. records do not stream from one stage to another. 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. 2006 134 of 179   Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The first block will be used by the upstream/producer stage to output data it is done with. BufferOp Selectively (being careful to to avoid deadlocks) disable buffering. ODBC Enterprise.4 Understanding Buffering There are two types of buffering in Enterprise Edition: ‘inter-operator transport’ and ‘deadlock prevention’. then setting APT_DEFAULT_TRANSPORT_BLOCK_SIZE to a multiple of (or equal to) the record size will resolve the problem. strictly speaking. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. Aggregator. stored in a retrieval system. If necessary. It usually only comes into play when the size of a single record exceeds the default size of the transport block.Information Integration Solutions Center of Excellence   Minimize non-combinable stages (as outlined in the previous section) such as Join. The second block will be used by the downstream/consumer stage to obtain data that is ready for the next processing step. . The default value is 8192. Remember. Oracle Enterprise. Each pair of operators that have a producer/consumer relationship will share at least 2 of these blocks.) 12. transmitted. Funnel. 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. All rights reserved. Once the upstream block is full and the downstream block is empty. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MAX_TRANSPORT_BLOCK_SIZE APT_MAX_TRANSPORT_BLOCK_SIZE July 17. so setting this value too high will result in a large amount of memory consumption. They are actually transferred in blocks (just like on old magnetic tapes) called “Transport Blocks”.1 Inter-Operator Transport Buffering Though it may look like it from the performance statistics and documentation might discuss ‘record streaming’. APT_MIN_TRANSPORT_BLOCK_SIZE o Specifies the minimum allowable block size for transferring data between players. 12. No part of this publication may be reproduced. This type of buffering (or ‘Record Blocking’) is rarely tuned. the value provided by a user for this variable is rounded up to the operating system's nearest page size. the blocks are swapped and the process begins again. (Buffering is discussed in more detail in the following section.

“Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job. Here is an example job fragment: Figure 55: Fork-Join example Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. Default is 1048576. No part of this publication may be reproduced. the framework calculates the block size for transferring data between players according to this algorithm: if (recordSize * APT_LATENCY_COEFFICIENT < APT_MIN_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MIN_TRANSPORT_BLOCK_SIZE else if (recordSize * APT_LATENCY_COEFFICIENT > APT_MAX_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MAX_TRANSPORT_BLOCK_SIZE else blockSize = recordSize * APT_LATENCY_COEFFICIENT APT_LATENCY_COEFFICIENT o Specifies the number of records to be written to each transport block APT_SHARED_MEMORY_BUFFERS o Specifies the number of Transport Blocks between a pair of operators. cannot be greater than 1048576. transcribed. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE. . APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. or translated into any language in any form by any means without the written permission of IBM. 12. APT_LATENCY_COEFFICIENT.4. transmitted. 2006 135 of 179 © 2006 IBM Information Integration Solutions. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records.2 Deadlock Prevention Buffering The other type of buffering. Cannot be less than 8192. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players. stored in a retrieval system.

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

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

Additionally. All rights reserved. use of this setting is only recommend for advanced users! FORCE_BUFFERING can be used to reveal bottlenecks in a job design during development and performance tuning. transmitted.Information Integration Solutions Center of Excellence operator will free run (up to a point) even when it has to write data to disk. and write increment can be set on a per-stage basis from the Input/ Advanced tab of the stage properties. buffer size. When it is not defined or defined to be the null string. as shown in the illustration below: Aside from ensuring that no dead-lock occurs. . This allows the job to run at the highest rate possible even when a Parallel Framework Red Book: Data Flow Job Design July 17. Default is 3 MB. No part of this publication may be reproduced. APT_BUFFERING_POLICY o Specifies the buffering policy for the entire job.  APT_BUFFER_MAXIMUM_MEMORY o Maximum memory consumed by each buffer operator for data storage. the default buffering policy is AUTOMATIC_BUFFERING. 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. queue bound. transcribed. BufferOps also have the effect of “smoothing out” production/consumption spikes. buffer free run. 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. but will almost certainly degrade performance and therefore shouldn’t be used in production job runs. the buffer mode. stored in a retrieval system.

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

etc) and directories where files are stored. intermediate work files. No part of this publication may be reproduced. as well as Annotation fields.5: Job Parameters. All DataStage jobs should be documented with Short Description field. Development Guidelines Modular development techniques should be used to maximize re-use of DataStage jobs and components. DataStage Template jobs should be created with:  standard parameters (for example. transmitted.5. 1. or translated into any language in any form by any means without the written permission of IBM.1:Directory Structures. stage names.  Naming conventions. An example directory naming structure is given in Section 2. transcribed. source and target file paths. and links. . This can also be used for integration with source code control systems.Information Integration Solutions Center of Excellence Appendix A: Standard Practices Summary This Appendix summarizes Standard Practices recommendations outlined in this document. 2006 140 of 179 © 2006 IBM Information Integration Solutions. and so forth. database login settings. 2.4: Working with Source Code Control Systems. file names. See Section 2.  Where possible. along with cross-references for more detail. using the Manager DSX export capability. 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.1 Environment Variable Settings)  annotation blocks Job Parameters should always be used for file paths. See Section 2. these standard parameters and settings should be made part of a Designer Job Template.3: Documentation and Annotation. It is the DataStage developer’s responsibility to make personal backups of their work on their local workstation. stored in a retrieval system. user. password. An example DataStage naming structure is given in Section 2. Standards It is important to establish and follow consistent standards in:  Directory structures for install and application support directories.2: Naming Conventions.  Create a standard directory structure outside of the DataStage project directory for source and target files. Parallel Framework Red Book: Data Flow Job Design July 17. To ease re-use. The scope of a parameter is discussed further in Section 3. especially for DataStage Project categories. database login properties…)  environment variables and their default settings (as outlined in Section 2. create re-usable components such as parallel shared containers to encapsulate frequently-used logic. All rights reserved. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously.  A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN.

or translated into any language in any form by any means without the written permission of IBM.7:Error and Reject Record Handling. while minimizing overhead. 2006 141 of 179 © 2006 IBM Information Integration Solutions. Leverage default type conversions using the Copy stage or across the Output mapping tab of other stages. . as outlined in Section 4:DataStage Data Types. using RCP to maximize re-use.  Use the parallel Transformer stage (not the BASIC Transformer) instead of the Filter or Switch stages. 4.  Use the Copy stage as a placeholder for iterative design. 3. 5.   DataStage Data Types Be aware of the mapping between DataStage (SQL) data types and the internal DS/EE data types. BASIC Routines are appropriate only for job control sequences. stored in a retrieval system. Server Shared Containers) within a parallel job.8: Component Usage. 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.  Always use parallel Data Sets for intermediate storage between jobs. minimizing overall run time. No part of this publication may be reproduced.  Use BuildOp stages only when logic cannot be implemented in the parallel Transformer. and to facilitate default type conversions. Objective 2: The partition method must match the business requirements and stage functional requirements. Further details are provided in Section 3. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. Standardized Error Handling routines should be followed to capture errors and rejects.Information Integration Solutions Center of Excellence Parallel Shared Containers should be used to encapsulate frequently-used logic. transcribed. This ensures that the processing workload is evenly balanced. 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. 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. transmitted. Component Usage As discussed in Section 3.

This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. 2006 142 of 179 © 2006 IBM Information Integration Solutions. Using the above objectives as a guide. stored in a retrieval system. Objective 3: Unless partition distribution is highly skewed. . especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. No part of this publication may be reproduced. Change Capture. transmitted. or translated into any language in any form by any means without the written permission of IBM. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. 6. Across jobs. It may also be necessary for Transformers and BuildOps that process groups of related records. Remove Duplicates.Information Integration Solutions Center of Excellence This includes. transcribed. All rights reserved. Change Apply. Merge. 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. use Auto partitioning (the default) Parallel Framework Red Book: Data Flow Job Design July 17. Note that in satisfying the requirements of this second objective. but is not limited to: Aggregator. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. 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. and Sort stages. minimize repartitioning. 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. Collecting Data Given the options for collecting data into a sequential stream. Join. This is particularly useful if downstream jobs are run with the same degree of parallelism (configuration file) and require the same partition and sort order. persistent Data Sets can be used to retain the partitioning and sort order.

use Sort Merge collector to produce a single. stored in a retrieval system. the following methodology should be applied when sorting in a DataStage Enterprise Edition data flow: a) Start with a link sort b) Specify only necessary key column(s) c) Don’t use Stable Sort unless needed d) Use a stand-alone Sort stage instead of a Link sort for options that not available on a Link sort:  Sort Key Mode. the 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. 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.Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel. as long as the Data Set has not been repartitioned or reduced. Merge. sequential ordered result set use a parallel Sort and a Sort Merge collector 8. 7. Sorting Using the rules and behavior outlined in Section 6: Sorting.1: Transformer NULL Handling and Reject Link.  Stage-Specific Guidelines As discussed in Section 8. Create Key Change Column. Join vs. See Section 9. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. or translated into any language in any form by any means without the written permission of IBM. use the Join or Merge stage. The Lookup stage is most appropriate when reference data is small enough to fit into available memory.  Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved.1: Lookup vs. . Create Cluster Key Change Column. transmitted.1. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. No part of this publication may be reproduced. If the Data Sets are larger than available memory resources. o Always place a reject link on a parallel Transformer to capture / audit possible rejects. 2006 143 of 179 © 2006 IBM Information Integration Solutions. transcribed. Output Statistics  Always specify “DataStage” Sort Utility for standalone Sort stages  Use the “Sort Key Mode=Don’t Sort (Previously Sorted)” to resort a sub-grouping of a previously-sorted input Data Set e) Be aware of automatically-inserted sorts  Set $APT_SORT_INSERTION_CHECK_ONLY to verify but not establish required sort order f) Minimize the use of sorts within a job flow g) To generate a single.

it is best to implement business rules natively using DataStage parallel components. 2006 144 of 179 © 2006 IBM Information Integration Solutions. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. 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. No part of this publication may be reproduced. use a SQL where clause to limit the number of rows sent to a DataStage job. When using Oracle. or Informix databases. Check the Director log for warnings. For maximum scalability and parallel performance. 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.1. A Sort method Aggregator should be used when the number of distinct key values is large or unknown. the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records. 9. .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. DB2. transcribed. use orchdbutil to properly import design metadata.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.  Database Stage Guidelines Where possible. All rights reserved. (See Section 9. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow. as discussed in Section 10. If possible. Use Hash method Aggregators only when the number of distinct key column values is small. which may indicate an underlying problem or data type conversion issue. 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. 10.2Capturing Unmatched Records from a Join). or when exception processing. stored in a retrieval system. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. use the native parallel database stages for maximum performance and scalability. transmitted.

and examine the job score by following the guidelines outlined in Section 12.4. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. 2006 145 of 179 © 2006 IBM Information Integration Solutions. . Parallel Framework Red Book: Data Flow Job Design July 17. Set the environment variable $OSH_PRINT_SCHEMAS to capture actual runtime schema to the Director log. Enable $APT_DUMP_SCORE by default. No part of this publication may be reproduced. error and warning messages from a running DS/EE job. transcribed. Set $DS_PX_DEBUG if the schema record is too large to capture in a Director log entry. stored in a retrieval system. transmitted.2 Understanding the Parallel Job Score.Information Integration Solutions Center of Excellence    The environment variable $DS_PX_DEBUG can be used to capture all generated OSH.

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

All rights reserved. transcribed. 2006 147 of 179 © 2006 IBM Information Integration Solutions. transmitted. stored in a retrieval system.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. . 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.

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

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

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

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

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

Parallel Framework Red Book: Data Flow Job Design July 17.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries. if you have a transfer that transfers all fields from input to output. 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: .add (# nullable fields)/8 for null indicators . the method APT_Record::estimateFinalOutputSize() can give you an estimate for a given record schema.Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. 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. . stored in a retrieval system. 2006 153 of 179 © 2006 IBM Information Integration Solutions. transmitted. As can APT_Transfer::getTransferBufferSize(). or translated into any language in any form by any means without the written permission of IBM. All rights reserved.one byte per column for field alignment (worst case is 3. No part of this publication may be reproduced. transcribed.

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

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

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

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.

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

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

Minimizing Record Movement for Maximizing Performance Now we have covered the basic rules and mechanics for hash-partitioning and sorting. but a different number of partitions. stored in a retrieval system. transmitted. or translated into any language in any form by any means without the written permission of IBM. July 17. 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. 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. let’s look at how we can capitalize on these behaviors for performance benefits. . Here is another possible outcome: Also: Consider the result of running the same job with the same data. No part of this publication may be reproduced. transcribed. All rights reserved.Information Integration Solutions Center of Excellence 3 102 Example: 6 rows in 2 partitions. 6 There is an exception to this rule: If your hash key has only one value. 2006 162 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.

transcribed. thereby increasing the value of this exercise. per store for all stores in the nation. No part of this publication may be reproduced. .Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets nation-wide. 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. as well as the average transaction amount per-item. transmitted. or translated into any language in any form by any means without the written permission of IBM. 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. They wish to determine the weighted average transaction amount peritem nation-wide. 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. 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. stored in a retrieval system. 2006 163 of 179 © 2006 IBM Information Integration Solutions. and append these values to the original data. All rights reserved.

and ‘Transaction Date’. Here. If you allow this environment variable to exist with any value. transcribed. 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’. automatic sort insertion. To do this. APT_NO_PART_INSERTION is NOT defined in your environment. 7 To enable automatic partition insertion. both on the output link of JoinSourceToAggregator_1): NOTE: In this job. Enterprise Edition will hash-partition and sort on ‘Item ID’. we will need to join each aggregator’s output back to the original data in order to get the original row count. transmitted. . Since the aggregator reduces row count (to the group count). not store). If you allow this environment variable to exist with any value. ‘Item ID’. and calculate the average of the ‘Transaction Amt’ column and place the results in a column named ‘Store Average Item Transaction Amt’. stored in a retrieval system. with ANY value. you will disable this facility. All rights reserved. ensure that the environment variable. To enable automatic sort insertion. Here you want to let DS/EE choose where to insert sorts and partitioners for you. with ANY value. The Aggregator StoreAverageItemTransactionAmt will aggregate the data on ‘Store ID’. The Aggregator stage NationalAverageItemTransactionAmt will aggregate the data on ‘Item ID’. 2006 164 of 179 © 2006 IBM Information Integration Solutions. to get the original data with the averages appended. there is only one sort and one repartition in the diagram. Enterprise Edition will hash-partition and sort on ‘Store ID’. This is the nation-wide transaction average per item (weighted by transaction. and automatic partition insertion must be turned on7. ‘Item ID’. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence appropriate places notice. This is the per-store transaction average per item. and ‘Transaction Date’ calculating the average of the ‘TransactionAmt’ column and place the results in a column named ‘National Average Item Transaction Amt’. ensure that the environment variable. Parallel Framework Red Book: Data Flow Job Design July 17. and ‘Transaction Date’. so you want to leave them enabled (the default). This is done with JoinSourceToAggregator_1 and JoinSourceToAggregator_2.

you will need to set the Aggregator’s “Method” to Sort. however. The hash method only requires the input data to be hashed. or translated into any language in any form by any means without the written permission of IBM.5 180.5 23 23 23 23 26. It does this by keeping running totals in memory for the aggregation for each output group.5 35.5 46 46 6.5 50.5 310. transmitted. 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.25 16. in return. All rights reserved. 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. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 26. not Hash.25 26. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation. The sort method requires the input data to be hashed and sorted.5 39.5 6.5 Store Average Item Transaction Amt 28 28 43 43 6. it does not guarantee output order.5 3.Information Integration Solutions Center of Excellence NOTE: For this scenario. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 39.25 26.5 6. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’.5 35.5 16.5 3. No part of this publication may be reproduced.25 180.5 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 56 23 85 7 6 7 3 98 87 534 PeekFinalOutput.5 180.5 180.5 35. transcribed. 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 50.25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17. stored in a retrieval system.5 310. . Therefore.5 3.

appended to the end of this document for masochistsiv) might suggest: 4. and 3.25 16.75 38. even sequential job execution does not alter the answer for the sort-related questions.25 38. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition. An examination of the job above would suggest: 6.5 44 44 Since both the Aggregator and Join expect the data to arrive hashed and sorted on the grouping key(s) —both operations that consume large amounts of CPU—a couple of questions arise with respect to efficiency: What is the minimum number of hash-partitioners needed to implement this solution correctly? What is the minimum number of sorts needed to implement this solution? What is the minimum number of times that sort will need to buffer the entire Data Set to disk to implement this solution? Though running the job sequentially eliminates questions related to partitioners.75 38.5 27 27 61. 2006 166 of 179 © 2006 IBM Information Integration Solutions. as only partition concurrency is affected by sequential execution.5 61. and 1. transmitted. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system.75 26. transcribed.75 38.25 44.25 44. A much better answer is: 1. A deeper examination (of the score dump. 6. more efficientv. All rights reserved. This is certainly an improvement on the previous answer. and 6. 3.e. i.Information Integration Solutions Center of Excellence 2 2 1 1 2 2 1 1 2 2 1 1 2 2 2 2 2 2 2 2 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 8 45 45 9 78 45 2 65 65 23 16. Here’s a screen shot of this. .5 33. 3. Parallel Framework Red Book: Data Flow Job Design July 17.25 44.: record adjacency assumes partition concurrency8.25 44.5 33.

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

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

Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt.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.75 38. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. .5 33. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. transcribed.5 3.5 26. No part of this publication may be reproduced.25 44. This sort will look exactly like the other sort stage: Remember to disable the ‘Stable Sort’ option if you do not need it (it will try to maintain row order except as needed to perform the sort. stored in a retrieval system.75 38.75 Store Average Item Transaction Amt 2 2 6 6 5 5 26.25 44.5 16.5 3.5 16.5 61. Output from above solution: Data Set 3: PeekFinalOutput.25 44. useful for preserving previous sort orderings).25 38.25 16.25 3.5 33. or translated into any language in any form by any means without the written permission of IBM. All rights reserved.25 44.5 27 27 61.75 38.25 16. as it is much more expensive than non-stable sorts. 2006 169 of 179 © 2006 IBM Information Integration Solutions. and it is enabled by default.

5 39. 000.000.5 6.Information Integration Solutions Center of Excellence Store Location Item ID Transaction Date Transaction Amt 1 1 2 2 1 1 1 1 2 2 1 1 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1 1 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 7 85 7 6 3 98 23 56 534 87 National Average Item Transaction Amt 35.25 180.5 50.5 310.5 23 23 180. No part of this publication may be reproduced. That is a LOT of saved processing power.25 26.5 39.000 records.5 23 23 26. and two sub-sorts—a much more efficient solution for large data volumes.000.000 records. Imagine a job with 100 million records as the input. 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 17.000 records.25 26. With the initial solution. a savings of 400.000.5 180. transmitted. The second solution only sorts (on disk) 100.000 records in addition to hashing 300. It looks very similar to the first solution. stored in a retrieval system. There is an even more efficient solution.000.5 35. .5 This solution produces the same result but is achieved with only one complete sort.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume.5 180. All rights reserved.5 35.5 310. a single partitioner. 100. we had to sort (on disk) 300.5 Store Average Item Transaction 28 28 43 43 6. but there is a critical difference.5 46 46 6.25 26. transcribed.5 50. 2006 170 of 179 © 2006 IBM Information Integration Solutions. and only hashes.5 35.5 6.

This is a potentially huge savings on large data volumes (remember the previous example). Data Set 4: PeekFinalOutput.25 16.25 16. stored in a retrieval system. Looks a lot like solution 1.a significant savings. the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed. transcribed. Comparing the efficiency of this solution with that of number two.5 26. automatic sort insertion.5 16. All rights reserved. for the same reasons.5 3.Information Integration Solutions Center of Excellence NOTE: In this job. 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.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26. except w/o the sort on the output of JoinSourceToAggregator. we have chosen to use the StoreLocation column as a part of our sorting key. The difference is on DistributeCopiesOfSourceData: Here.25 44. and automatic partition insertion must be turned off vii. No part of this publication may be reproduced. However it will not create additional processes to handle the records and re-order them. This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator. transmitted. Also.5 3.5 27 27 July 17.25 16.25 44. but NOT to use it for hashing.5 3. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. or translated into any language in any form by any means without the written permission of IBM. . Here is the output from this version of the job. we saved a sub-sort on 100 million records .

5 35.5 33.5 46 46 6.5 6.5 310.5 180. 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. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35. in addition to the heavy penalty paid in disk I/O for using a full sort.25 26.25 44.5 Store Average Item Transaction Amt 28 28 43 43 6.5 33.25 38.5 50.5 39.75 38. 2006 172 of 179 © 2006 IBM Information Integration Solutions.5 180.5 50.75 61.5 35. Here is a screen shot of a sort running on 40 million records: Parallel Framework Red Book: Data Flow Job Design July 17.5 180. sort. stored in a retrieval system.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. transcribed.5 35. by definition. transmitted.5 6.25 180.75 38.5 39.5 23 23 23 23 26.75 38. No part of this publication may be reproduced. .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.5 44 44 PeekFinalOutput.5 310. All rights reserved.25 26. or translated into any language in any form by any means without the written permission of IBM.25 26.

This allows down-stream stages to be processing data during the sorting process. 2006 173 of 179 © 2006 IBM Information Integration Solutions. the link sort in solution 2. requiring a large amount of scratch diskx. transmitted. Here is an example of a ‘sub-sort’: Here. All rights reserved. and the link sort in solution three. .Information Integration Solutions Center of Excellence As you can see. 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. although ~5 million records have entered the sort. stored in a retrieval system. This is because a standard sort requires all rows to be present in order to release the first row. no rows have left yet. instead of waiting until all 40 million records have been sorted (in this instance. we are sub-sorting the data we sorted in the previous diagram). you can clearly see that a sub-sort does not inhibit pipe-lining--very nearly the same number of rows have entered and left the sort stage (and NO buffering is required to perform the subsort). This situation is analogous to all of the sorts in solution 1. transcribed.

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

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

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

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

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

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

Sign up to vote on this title
UsefulNot useful