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.

...........39 3............................................................................................................................................................................................................................................................................................................................................................3 JOB DESIGN TEMPLATES....................................................................................81 7............ stored in a retrieval system.............................................................76 6................................................................................................................. transcribed.................... ..18 2...................................................................................................................................... transmitted.....2 JOB TYPES....2 NULL HANDLING..........................81 7..............7 1.......................................................................................................................82 7.......................................................79 7 FILE STAGE USAGE...............................................................................................................................................................................1 PARTITION AND SORT KEYS.............................................................................................................................................................74 6..................................73 6 SORTING.....................................................................................................................................74 6..................................................................2 MONITORING PARTITIONS.......................................85 8 TRANSFORMATION LANGUAGES...............................................................................................77 6...............................78 6.....................31 2................7 ERROR AND REJECT RECORD HANDLING................................................2 NAMING CONVENTIONS...........................................39 3........................................................................................................................................29 2..........7 SORT METHODOLOGY.......................................................... No part of this publication may be reproduced.................................................................................................1 DIRECTORY STRUCTURES......56 4.........5 UNDERSTANDING A JOB’S ENVIRONMENT....................................6 PARALLEL SHARED CONTAINERS..........1 JOB SEQUENCES.................................5 COLLECTOR TYPES...................................................68 5.58 5 PARTITIONING AND COLLECTING......................................................................................................1 WHICH FILE STAGE TO USE......................................Information Integration Solutions Center of Excellence Table of Contents 1 DATA INTEGRATION OVERVIEW.....................................................4 DEFAULT JOB DESIGN..87 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...................................................................................................................................................................................................... 2006 4 of 179 © 2006 IBM Information Integration Solutions.........................................................4 WORKING WITH SOURCE CODE CONTROL SYSTEMS......................54 4....................3 LINK SORT AND SORT STAGE...........13 2..................................................................................................................................................................................................................41 3.....................................................................................................8 TUNING SORT............................................................................67 5................35 3 DEVELOPMENT GUIDELINES...................................................................................77 6.....................................................................................................................................................................................................43 3..................................................................................................70 5................................................................................................................4 STABLE SORT............................................................................................40 3.............3 PARTITION METHODOLOGY......................................... All rights reserved.................................................................................2 DATA SET USAGE.........................................................................................................43 3.3 RUNTIME COLUMN PROPAGATION......................................75 6......................................................................................39 3...............................................................................6 COLLECTING METHODOLOGY..........................................59 5......3 DOCUMENTATION AND ANNOTATION.....................................................................................................................................................................................................................2 COMPLETE (TOTAL) SORT..............................................................................................................................................................................................................................................................1 PARTITION TYPES...59 5..............................................................................................................................72 5........................................1 MODULAR DEVELOPMENT .......................................................51 4 DATASTAGE DATA TYPES.........42 3.....................................................................................................................................................2 ESTABLISHING JOB BOUNDARIES.........................................................79 6..................5 JOB PARAMETERS............4 PARTITIONING EXAMPLES...........................................4 COMPLEX FLAT FILE STAGE....8 COMPONENT USAGE...............................................................................................................................................................................5 SUB-SORTS........................13 2......................................................................6 1.....................................................................................................................................................................................3 SEQUENTIAL FILE STAGES (IMPORT AND EXPORT).........................................................................................................................................................................................................................................81 7.................................................................................6 AUTOMATICALLY-INSERTED SORTS...........................................................................................................................................................................................8 2 STANDARDS........

.. transmitted..................................................................................................................148 APPENDIX D: ESTIMATING THE SIZE OF A PARALLEL DATA SET...........................1 WARNING ON SINGLE-NODE CONFIGURATION FILES...124 11..................................................................2 MODIFY STAGE................................................................................................................................................................................................87 8.....4 ODBC ENTERPRISE GUIDELINES.......................................................................120 11 TROUBLESHOOTING AND MONITORING........3 HOW TO ISOLATE AND DEBUG A PARALLEL JOB................................................................................................140 APPENDIX B: DATASTAGE NAMING REFERENCE....................................................................................................................................................................................................................3 INFORMIX DATABASE GUIDELINES.........................................................1 LOOKUP VS............................................................................................................................................................................................................................................................................................................... MERGE.........................................................117 10....................5 ORACLE DATABASE GUIDELINES............................... JOIN VS....................5 INTERPRETING THE PARALLEL JOB SCORE............................................................119 10.......1 DATABASE DEVELOPMENT OVERVIEW................................................... transcribed.........2 UNDERSTANDING OPERATOR COMBINATION...........................153 APPENDIX E: ENVIRONMENT VARIABLE REFERENCE.........................94 9.......129 12.........................................................7 TERADATA DATABASE GUIDELINES............................................................129 12...........................................Information Integration Solutions Center of Excellence 8.......94 9..............4 VIEWING THE GENERATED OSH.................. All rights reserved.................................146 APPENDIX C: UNDERSTANDING THE PARALLEL JOB SCORE.................................................................................................................91 9 COMBINING DATA.....................................127 12 PERFORMANCE TUNING JOB DESIGNS.......................................................................................................................................................................................2 DEBUGGING ENVIRONMENT VARIABLES................. No part of this publication may be reproduced..............................................................................................................133 12...113 10..............................................................................................1 TRANSFORMER STAGE...........................................................................................................2 CAPTURING UNMATCHED RECORDS FROM A JOIN................................... 2006 5 of 179 © 2006 IBM Information Integration Solutions................................................... ....95 10 DATABASE STAGE GUIDELINES.................................................................................................................. stored in a retrieval system.........................................................134 APPENDIX A: STANDARD PRACTICES SUMMARY..............154 APPENDIX F: SORTING AND HASHING ADVANCED EXAMPLE...........125 11.....126 11........................................................................3 MINIMIZING RUNTIME PROCESSES AND RESOURCE REQUIREMENTS.....................................................................................................96 10..........................131 12..............................4 UNDERSTANDING BUFFERING...............................................................................94 9..............124 11.114 10..............................96 10...................................................................................................................................103 10................................................ or translated into any language in any form by any means without the written permission of IBM.........1 HOW TO DESIGN A JOB FOR OPTIMAL PERFORMANCE...........................................................................124 11..............................................................................2 DB2 GUIDELINES................................................................160 Parallel Framework Red Book: Data Flow Job Design July 17..........3 THE AGGREGATOR STAGE...........6 SYBASE ENTERPRISE GUIDELINES....................................................................................................................................................

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

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

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

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

or translated into any language in any form by any means without the written permission of IBM. transcribed. stored in a retrieval system. transmitted. 2006 10 of 179 © 2006 IBM Information Integration Solutions.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. All rights reserved. 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. and the write-through cache records from the last load of the target database are merged. the validated records.2 Hybrid Jobs The following example hybrid job demonstrates several interesting techniques that might be used in more complex jobs. In this case. transcribed. we replicated the Oracle structure (lower input) for each country found in the write-through cache country dataset (upper input). This job also loads the target database table and creates write-through cache. The non-validated records. producing an ordered and linked associative table. All rights reserved. forcing other jobs that might depend on this data to access the existing (not updated) target database table. If the key column is NULL. it is rejected by the transformer to a reject port and the validation is not performed for those records. 2006 11 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. The key column for a Referential Integrity check is validated by a Transformer stage.2. 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.Information Integration Solutions Center of Excellence 1. . The merged records are grouped and ordered before being de-duplicated to remove obsolete records. In this case. transmitted. 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). or translated into any language in any form by any means without the written permission of IBM. Some of the more interesting solutions in this job are circled. Parallel Framework Red Book: Data Flow Job Design July 17. The de-duplicated records are re-grouped and ordered before calculation of the terminating keys. No part of this publication may be reproduced. if the load fails. the cache is deleted.

2006 12 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 1. . stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.2. 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.

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

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

data directories are implemented for each deployment phase of a job (development. stored in a retrieval system.. /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.. 2006 15 of 179 . All rights reserved. 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 . /Project_Z Within the separate Staging file system.. /Project_Z July 17. location of source data files. target data files. transcribed./archive of these development phases may be present on a local file not all system. 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 . system integration. files are separated by Project name as shown below.. Within each deployment directory.. transmitted. error and reject files.. .. If the file system is not /archive /archive /archive shared across multiple servers.. or translated into any language in any form by any means without the written permission of IBM. /Staging /dev Top-Level Directory development data tree. and production) as appropriate. qa. .

/Project_Z . All rights reserved. or translated into any language in any form by any means without the written permission of IBM. This will provide a convenient vehicle to group and manage resources used by a project. The directory structure will be made transparent to the DataStage application. A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project... Environment variables are a critical portability tool. which will enable DataStage applications to move through the life cycle without any code changes. 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.. transmitted. July 17. 2006 . . transcribed. . 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. such as the operating system./Project_Z publication may be reproduced.2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities. through the use of environment variables. 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...Information Integration Solutions Center of Excellence 2..1. another application or middle ware. or it could require scripts for example integrating with an Enterprise Scheduler. 16 of 179 ...

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

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

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

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

allowing related objects to be grouped together. which can provide impact analysis. DataStage enforces the top level Directory Structure for different types of Objects (for example. transmitted.2. The name of a DataStage Project may only be 18 characters in length.2. No part of this publication may be reproduced. as well as documentation and auditing. it can contain alpha-numeric characters and it can contain underscores. It may or may not have a one to one relationship with an organizations’ project of work. All rights reserved. transcribed. 2.3 Job Category Naming Within Designer.1 Projects Each DataStage Project is a standalone repository. This factor often can cause terminology issues especially in teamwork where both business and developers are involved. However with the limit of 18 characters the name is most often composed of abbreviations. Jobs. or translated into any language in any form by any means without the written permission of IBM.4.2. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17.2 Category Hierarchy DataStage organizes objects in its repository by Categories. dialog box fields that specify a new category have only one input area for defining the Category name. and Production (“Prod”). . For example. 2.4 Naming Conventions by Object Type 2. stored in a retrieval system. 2006 21 of 179 © 2006 IBM Information Integration Solutions. developers have the flexibility to define their own Directory or Category hierarchy. Table definitions…). are Alpha Numeric and can also contain both Spaces and Underscores. Category Names can be long. Shared Containers. 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. The suffix of a Project name should be used to identify Development (“Dev”).4. Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”).2. Below this level. Routines. Therefore Directory names should be Word Capitalized and separated by either an underscore or a space. Test (“Test”).Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage.4.

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

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

the TableDefs have been grouped into a master category of Custom. transmitted. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”. transcribed. No part of this publication may be reproduced.: Datasets. An alternative implementation is to set the “Data source name” to that of the source system or schema. Figure 7: Table Definition Categories 2. e. stored in a retrieval system.2.4. numbers. Because the name of can be long. Parallel Framework Red Book: Data Flow Job Design July 17. . This will be discussed further in Section 2.4. Each subject area will have a master category.2 Category Hierarchy. and underscores only.2. 2006 24 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence should be deleted as soon as possible. Job and Job Sequence names should be descriptive and should use word capitalization to make them readable. Examples of Job naming are: • CodeBlockAggregationJob • CodeBlockProcessingSeq Jobs should be organized under Category Directories to provide grouping such that a Directory should contain a Sequence Job and all the Jobs that are contained within that sequence.g: DWPH1 or ECRP. or translated into any language in any form by any means without the written permission of IBM. In this example. with sub-categories intended to identify the type of the source. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters. All rights reserved. e.g. Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”.

links represent the flow of a message from one activity / step to the next.4. the following Class Word naming is recommended: • Psc = Parallel (Enterprise Edition) Shared Container • Ssc = Server Edition Shared Container IMPORTANT: Use of Server Edition Shared Containers is discouraged within a parallel job. It is optional as to whether you decide to change this code to something meaningful.4. Within a Job Sequence. To differentiate between Parallel Shared Containers and Server Shared Containers. or translated into any language in any form by any means without the written permission of IBM. having a descriptive link name reduces the chance for errors (for example.2. All rights reserved.6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. stage editors identify links by name. during Link Ordering). It is particularly important to establish a consistent naming convention for link names. transcribed. transmitted. stored in a retrieval system. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. a character code is automatically added to that instance of its use throughout the project.8 Links Within a DataStage Job. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. instead of using the default “DSLink#” (where “#” is an assigned number). links are objects that represent the flow of data from one stage to the next. Furthermore. so word capitalization should be used for readability. No part of this publication may be reproduced. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2. 2006 25 of 179 © 2006 IBM Information Integration Solutions. 2.2. When a Shared Container is used. The class word suffix is “Parm”. . Within the graphical Designer environment. Therefore the parameter name must be made readable using Capitalized words separated by underscores.7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores.2.4. 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.Information Integration Solutions Center of Excellence 2. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17.

based on the order the object was added to the flow. Sequential File. stage names must be unique. 3.4. transcribed.2. through Job reporting). Transformer. Instead of using the full object name. after the subject name and subject modifier. 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. the same name may be appropriate for multiple links. 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.4.) • 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. 2006 26 of 179 © 2006 IBM Information Integration Solutions. . (The DataStage Designer does not require link names on different stages to be unique. the link name should include the link type (reference. No part of this publication may be reproduced. transmitted.Information Integration Solutions Center of Excellence example. 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. All rights reserved. always specify a unique link name within a particular Job or Job Sequence by including a number. These names are based on the type of stage (object) and a unique number. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12.2 DataStage Naming Reference.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. establishing standardized link names makes it easier to understand results and audit counts. a 2. etc) • The type of movement may optionally be part of the Class Word. Parallel Framework Red Book: Data Flow Job Design July 17. or 4 character abbreviation should be used for the Class Word suffix. or translated into any language in any form by any means without the written permission of IBM. Database. Within a Job or Job Sequence. In this case. Merge.

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.

RcR_GLOBAL_BUSN_CAT_TYP table from jpORACLE_SERVER using the ODBC driver. transcribed. transmitted. There are no selection criteria in the WHERE clause. 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. There are no selection criteria in the WHERE clause. Parallel Framework Red Book: Data Flow Job Design July 17. Oracle Enterprise stage read: Read the GLOBAL. . More complex operators or operations should have correspondingly longer and more complex explanations on this tab. or translated into any language in any form by any means without the written permission of IBM. ODBC Enterprise stage read: Read the GLO.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. stored in a retrieval system. All rights reserved. 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. Each stage should have a short description of its function specified within the stage properties. No part of this publication may be reproduced. 2006 30 of 179 © 2006 IBM Information Integration Solutions.

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

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

Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group.2 Category Hierarchy. establishing and following a consistent naming and categorization standard is essential to the change management process. the developer should create a local backup prior to implementing any extensive changes. Category. In either case. It can be done interactively by the developer or project manager using the Manager client. These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide. There is no server-side project export facility.4. or by the individual DataStage developer. The DataStage client includes Windows command-line utilities for automating the export process.2. July 17. Assigning related objects to the same category provides a balanced level of granularity when exporting and importing objects with external source code control systems.4. All rights reserved. 2. • Select the object or category in the Manager browser. 2. transmitted.4. stored in a retrieval system. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process. The DataStage Manager can export at the Project. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu. 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. as explained in this section. and individual Object levels. . All exports from the DataStage repository are performed on the Windows workstation.2 Using Object Categorization Standards As discussed in Section 2. transcribed.

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

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

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

4. but part of every job design so that it can be easily enabled for debugging purposes. Outputs actual runtime metadata (schema) to DataStage job log.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. 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. transcribed.5. 2006 37 of 179 © 2006 IBM Information Integration Solutions. or may be set on an individual basis within the properties for each job. stored in a retrieval system. All rights reserved. It may be helpful to create a Job Template and include these environment variables in the parameter settings. This setting should be disabled by default. 2. This variable should be included in all job parameters so that it can be easily changed at runtime. or translated into any language in any form by any means without the written permission of IBM. processes. Several environment variables are evaluated only for their presence in the environment (for example.5. This setting should be disabled by default. No part of this publication may be reproduced. The count is per operator per partition. Extremely useful for understanding how a job actually ran in the environment. but part of every job design so that it can be easily enabled for debugging purposes. transmitted. Does not report PIDs of DataStage “phantom” processes started by Server shared containers. 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 17.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. APT_SORT_INSERTION_CHECK_ONLY). and to document a few that are not part of the documentation. These settings can be made at the project level. Outputs EE score dump to the DataStage job log. (see 12.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment. This section is intended to call attention to some specific environment variables.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. and Data Sets. .1. providing detailed information about actual job flow including operators. $UNSET 2.

All rights reserved. transcribed.1A platforms only: On Tru64 platforms. The environment variables mentioned in this document are summarized in Appendix D: Environment Variable Reference. stored in a retrieval system. or changing the default behavior of specific Enterprise Edition stages. Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. . This setting instructs EE to use named pipes rather than shared memory for local data transport. or translated into any language in any form by any means without the written permission of IBM. a number of environment variables will be mentioned for tuning the performance of a particular job flow. 2006 38 of 179 © 2006 IBM Information Integration Solutions. you must define the environment variable $APT_IO_NOMAP On Tru64 5.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). transmitted.5. 2.2 Additional Environment Variable Settings Throughout this document. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide.1. assisting in debugging. the environment variable $APT_PM_NO_SHARED_MEMORY should be set to 1 to work around a performance issue with shared memory MMAP operations.

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

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.

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

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

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

Send the reject stream to an *. or translated into any language in any form by any means without the written permission of IBM.rej or *. Transformer Stage Option Funnel the reject stream back to the output stream(s).err extension when rejects can be ignored but need to be recorded.err file or tag and merge with the output stream. 3. Send the reject stream to a reject file and halt the job. however. transcribed.rej file or tag and merge with the output stream.rej extension is used when rejects require investigation after a job run. a reject link must exist on that stage. Satisfies requirements for a reject conditional output stream. To capture rejects from a target database. No part of this publication may be reproduced. The *. Alternatively. 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 extension. and Oracle Enterprise) offer an optional reject link that can be used to capture rows that cannot be written to the target database. The reject option should be used in all cases where active management of the rejects is required. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. 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. Alternatively. it must have a *.7.4 Reject Handling with target database stages Some database stages (such as DB2/UDB Enterprise. All rights reserved. transmitted. a shared container error handler can be used. reject rows will not be captured. Rejects should not exist but should not stop the job. Cannot satisfy requirements of any conditional output stream and is rejected by the default output stream. .Information Integration Solutions Center of Excellence 1. OR 2. Otherwise. If a file is created from the reject stream. Rejects should not exist and should stop the job. and be reviewed by the Data Steward. 2006 46 of 179 © 2006 IBM Information Integration Solutions. they only need to be recorded but not acted on. 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. ODBC Enterprise. Pass rows that fail to be written to the reject stream.rej file extension. Send the reject stream to an *. a shared container error handler is used. the *. it must have a *. Rejects can exist in the data.

rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504. The appropriate file type. Reject link exists. 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. Send the reject stream to a *. The job serial number (jpJOBSERIALNO) and a period “. and should be reviewed by the Data Steward. 3. transcribed. or translated into any language in any form by any means without the written permission of IBM. The Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job.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.Rejects are tracked by count only.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.7. The job name (jpJOB_NAME) and a underscore “_”. . All rights reserved. The project phase (jpENVIRON) and a underscore “_”. For example. 2. 4. The project name (jpPROJECT_NAME) and a underscore “_”. and 5. stored in a retrieval system.”. one of “rej” or “err”.rej file. Only records that match the given table definition and database constraints are written. 3.1 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. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. and gathered using a Funnel stage that feeds a Sequential File stage.7. transmitted. 2 Rejects should not exist but should not stop the job. job DECRP_N_XformClients in the ECR_FACTS project in the development environment with a serial number of 20060201-ETL-091504 would have these reject and error file names: ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504.5.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. This section deals with both methods of handling errors. No part of this publication may be reproduced.

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

. 2006 49 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence The input to the Column Export stage explicitly converts the data unique to the reject stream (in this case. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. 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. Track*) to a single output column. transcribed. transmitted.

. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream. Switch Transformer In this example. In either case. One or more columns may have been selected for replacement when a reference key is found. The following table identifies the tagging method to be used for the previously cited operators. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values. transcribed.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.7. A failed switch will reject an intact input row show key fails to resolve to one of the Switch output stream. Connect the reject port to a Transformer stage where columns are set to specific values. Connect the reject port to a Transformer stage where columns are set to specific values. All rights reserved. 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. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. 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. or translated into any language in any form by any means without the written permission of IBM. transmitted. 2006 50 of 179 © 2006 IBM Information Integration Solutions. 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. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3. 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 output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams.5. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key. OR with columns contain illegal values for some operation performed on said columns.

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

the Copy stage should be used as a placeholder so that new stages can be inserted easily should future requirements change. and to use other stages (such as Copy) when derivations are not needed. • The Copy stage should be used instead of a Transformer for simple operations including: . o Unless the Force property is set to “True”.8. 3. 3.2 Copy Stage For complex data flows. . This is discussed in more detail in Parallel Framework Standard Practices: Administration. transcribed.Dropping Columns July 17. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime. transmitted. stored in a retrieval system. Used in this manner. Enterprise Edition will optimize this out at runtime) . and necessary.Server shared containers Note that BASIC Routines are still appropriate. it is important to minimize the number of transformers. 3.4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. in sort order. For this reason.BASIC Transformers . they should not be used for long-term archive of source data.Renaming Columns .8. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run. 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. always write to parallel Data Sets. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. it can be used at the end of a data flow o For simple jobs with only two stages. and Production Automation.Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: .BASIC Routines . No part of this publication may be reproduced. Management.Job Design placeholder between stages (unless the Force option =true. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs. Data Sets offer parallel I/O on read and write operations. without overhead for format or data type conversions. NOTE: Because parallel Data Sets are platform and configuration-specific. and in Enterprise Edition native format. Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form.8. it is best to develop a job iteratively using the Copy stage as a “placeholder”. Since the Copy stage does not require an output link.3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs.

For example.8. drop (if Runtime Column Propagation is disabled). 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.1. See Section 8. Instead. 2006 53 of 179 © 2006 IBM Information Integration Solutions. transcribed. 3. • NEVER use the “BASIC Transformer” stage in large-volume job flows. and default type conversion can also be performed by the output mapping tab of any stage.5.3 Then B=”X” If A=4.6. if possible.5 BuildOp stages BuildOps should only be used when: . Parallel Framework Red Book: Data Flow Job Design July 17. • Because the parallel Transformer is compiled. stored in a retrieval system. implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. the derivation expression: If A=0. or translated into any language in any form by any means without the written permission of IBM.1. All rights reserved.Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4.7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B. No part of this publication may be reproduced. transmitted. . user-defined functions and routines can expand parallel Transformer capabilities.2: Default and Explicit Type Conversions] Note that rename. and character string trimming. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines. performance should this should be tested in isolation to identify specific cause of bottlenecks. Consider. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible. null handling. Even then its use should be restricted and the routines should be converted as soon as possible.2.Existing Transformers do not meet performance requirements As always.Complex reusable logic cannot be implemented using the Transformer or . • The Modify stage can be used for non-default type conversions.2: Modify Stage.

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

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

It allocates a single bit to mark a field as null.It designates a specific field value to indicate a null. transcribed.2 Null Handling DataStage Enterprise Edition represents nulls in two ways: . All rights reserved. When used in these stages. stored in a retrieval system. 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. the PadString Transformer function can be used to pad a variablelength (Varchar) string to a specified length using a specified pad character. depending on the source and result data types. Enterprise Edition pads the remaining length with NULL (ASCII zero) characters by default. 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. Note that PadString does not work with fixed-length (CHAR) string types. You must first convert a Char string type to a Varchar type before using PadString. • 4. Parallel Framework Red Book: Data Flow Job Design July 17. for example. When converting from variable-length to fixed-length strings. • The environment variable APT_STRING_PADCHAR can be used to change the default pad character from an ASCII NULL (0x0) to another character. Sequential File and DB2/UDB Enterprise targets) allow the pad character to be specified in their stage or column definition properties.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. Some stages (for example. This type of representation is called an out-of-band null. Enterprise Edition displays a warning message in the job log. the specified pad character will override the default for that stage only. This type of representation is called an in-band null. 2006 56 of 179 © 2006 IBM Information Integration Solutions. . transmitted. In these instances. 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. No part of this publication may be reproduced. an ASCII space (0x20) or a Unicode space (U+0020). for example a numeric field’s most negative possible value. . • As an alternate solution.

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

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

and provides guidelines for appropriate use in job designs. If the prior stage was running sequentially. Keyed partitioning is used when business rules (for example. it is important that this distribution does not violate business requirements for accurate data processing. 2006 59 of 179 © 2006 IBM Information Integration Solutions. transcribed. Join) require processing on groups of related records. 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.Information Integration Solutions Center of Excellence 5 Partitioning and Collecting Partition parallelism is a key to establishing scalable performance of DataStage Enterprise Edition. No part of this publication may be reproduced. 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. Remove Duplicates) or stage requirements (for example. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. Partitioners distribute rows of a single link into smaller segments that can be processed independently in parallel. All rights reserved. Keyed partitioning examines the data values in one or more key columns. as shown in this example: Stage running sequentially Figure 18: “fan-out” icon Collectors combine parallel partitions of a single link for sequential processing. . 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. Partitioning methods are separated into keyless and keyed classes: . a “fan-out” icon is drawn on the link within the Designer canvas. For this reason. Partitioners exist before any stage that is running in parallel.Keyless partitioning distributes rows without regard to the actual data values. Parallel Framework Red Book: Data Flow Job Design July 17.1 Partition Types While partitioning allows data to be distributed across multiple processes running in parallel. It also provides tips for monitoring jobs running in parallel. transmitted. Stage running sequentially Stage running in parallel 5.

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

transmitted.1 Same Partitioning Same partitioning in fact performs no partitioning to the input Data Set. Instead. 5. 5. a parallel Sort).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. as illustrated on the right: Same partitioning doesn’t move data between partitions (or. in the case of a cluster or Grid. or translated into any language in any form by any means without the written permission of IBM. and is appropriate when trying to preserve the grouping of a previous operation (for example. Distributes rows evenly across partitions in a random partition assignment. although it can be explicitly set or cleared in the “Advanced” stage properties of a given stage.1. 2006 61 of 179 © 2006 IBM Information Integration Solutions.2. transcribed. stored in a retrieval system. Each partition receives the entire Data Set. in a round robin partition assignment. if the Preserve Partitioning flag was set. for example). Because Same does not redistribute existing partitions. as shown in on the right: Figure 22: Preserve Partitioning option The Preserve Partitioning flag is part of the Data Set structure. between servers). 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. There are some cases when the input stage requirements prevent partitioning from being preserved. the degree of parallelism remains unchanged: Parallel Framework Red Book: Data Flow Job Design July 17.1. 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. Distributes rows evenly across partitions. All rights reserved. and its state is stored in persistent Data Sets.Information Integration Solutions Center of Excellence some stages (Sort. a warning will be placed in the Director log indicating that Enterprise Edition was unable to preserve partitioning for a specified stage. it retains the partitioning from the output of the upstream stage. In these instances. . No part of this publication may be reproduced.

1.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5. similar to dealing cards: Round robin partitioning has a fairly low overhead. Since the random partition number must be calculated. transcribed. stored in a retrieval system. .1. or translated into any language in any form by any means without the written permission of IBM. it is rarely used in real-world data flows. Entire partitioning may have a performance impact. 3 2 1 0 .3 Random Partitioning Like Round Robin. 2006 62 of 179 © 2006 IBM Information Integration Solutions. regardless of the current $APT_CONFIG_FILE 5. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). As a result. the downstream stage runs with the degree of parallelism used to create the Data Set. the order that rows are assigned to a particular partition will differ between job runs. Random partitioning has a slightly higher overhead than Round Robin partitioning. as the complete Data Set must be distributed across the network to each node. 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17. 5. .2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment. . 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). 3 2 1 0 . .2. but using a random assignment. All rights reserved. While in theory Random partitioning is not subject to regular data patterns that might exist in the source data. transmitted. Random partitioning evenly distributes rows across partitions.2. On clustered and Grid implementations. …8 7 6 5 4 3 2 1 0 ENTIRE .2.1.4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition. Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning. No part of this publication may be reproduced. Since optimal parallel processing occurs when all partitions have the same workload.

2006 63 of 179 © 2006 IBM Information Integration Solutions. Keyed partitioning is used when business rules (for example. If the source data values are evenly distributed within these key column(s). transmitted. All rights reserved. ensuring that records with the same values in those key column(s) are assigned to the same partition.Information Integration Solutions Center of Excellence 5.1. No part of this publication may be reproduced. and there are a large number of unique values. Join) require processing on groups of related records. 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. consider the following sample Data Set: ID 1 2 3 4 5 6 7 8 9 10 LName Ford Ford Ford Ford Dodge Dodge Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Horace John Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 17840 Jefferson 75 Boston Boulevard 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Values of key column …0 3 2 1 0 2 3 2 1 1 HASH 0 3 0 3 1 1 1 2 2 2 Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. then the resulting partitions will be of relatively equal size.1. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. 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. . As an example of hashing. 5.3. 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.3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns.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. stored in a retrieval system. transcribed.

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 Hashing on key column LName would produce the following results: Partition 0: ID 5 6 LName Dodge Dodge FName Horace John Address 17840 Jefferson 75 Boston Boulevard Partition 1: ID 1 2 3 4 7 8 9 10 LName Ford Ford Ford Ford Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. transmitted. stored in a retrieval system. . No part of this publication may be reproduced. 2006 64 of 179 © 2006 IBM Information Integration Solutions.

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

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

and each stage is processing an equal number (12.2 Monitoring Partitions At runtime. and select the “Show Instances” option. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. 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. databases) c) and. as shown in the example log below where the stage named “Peek” is running with four degrees of parallelism (partition numbers zero through 3): Figure 24: Partition numbers as shown in Director log To display row counts per partition in the Director Job Monitor window. The partition number is appended to the stage name for messages written to the Director log. This is very useful in determining the distribution across parallel partitions (skew). No part of this publication may be reproduced. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17.500) of rows for an optimal balanced workload. Specific details on interpreting the parallel job score can be found in 12. All rights reserved. transcribed. stored in a retrieval system. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True.4. as illustrated below. Partitions are assigned numbers. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name). transmitted. in some cases. 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. In this instance. . 2006 67 of 179 © 2006 IBM Information Integration Solutions. right-click anywhere in the window.Information Integration Solutions Center of Excellence 5. if specified. or translated into any language in any form by any means without the written permission of IBM.2Understanding the Parallel Job Score. starting at zero.

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

assigning related records to the same partition if required Any stage that processes groups of related records (generally using one or more key columns) must be partitioned using a keyed partition method. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). o Across jobs. but is not limited to: Aggregator. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. transmitted. 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.Information Integration Solutions Center of Excellence Objective 2: The partition method must match the business requirements and stage functional requirements. and Sort stages. All rights reserved. transcribed. Note that in satisfying the requirements of this second objective. minimize repartitioning. Parallel Framework Red Book: Data Flow Job Design July 17. Change Capture. 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. stored in a retrieval system. No part of this publication may be reproduced. Change Apply. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. Join. . it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. 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. 2006 69 of 179 © 2006 IBM Information Integration Solutions. It may also be necessary for Transformers and BuildOps that process groups of related records. Using the above objectives as a guide. Objective 3: Unless partition distribution is highly skewed. or translated into any language in any form by any means without the written permission of IBM. Remove Duplicates. Merge. This includes. 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 apply Same partitioning to the downstream links. To add aggregate columns to every detail row.2Sorting and Hashing Advanced Example.Information Integration Solutions Center of Excellence 5. 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.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. a Transformer is used to extract data from a single header row of an input file. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator. the partitioning and sorting of this scenario can be optimized. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly. or through Auto partitioning): Parallel Framework Red Book: Data Flow Job Design July 17. 5. Using a “standard” solution.4.4. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. 5.4 Partitioning Examples In this section. Within the Transformer. The output of the Aggregator is then sent to the second input of the Join. 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. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row. more advanced partitioning and sorting examples are given in 12. . Additional. transmitted. No part of this publication may be reproduced. we can move the Hash partition and Sort before the Copy stage. 2006 70 of 179 © 2006 IBM Information Integration Solutions. Because the Join and Aggregator use the same partition keys and sort order. we’ll apply the partitioning methodology defined earlier to several example job flows. a new output column is defined on the header and detail links using a single constant value derivation.4.2 Partitioning Example 2 – Use of Entire Partitioning In this example. stored in a retrieval system. transcribed. on closer inspection.

or translated into any language in any form by any means without the written permission of IBM. Remembering that the degree of parallel operation is limited by the number of distinct values. 2006 71 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions . Although functionally correct. the link order of the Inner Join is significant. 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). For this reason. 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. there is no need to pre-sort the input to the Join.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. stored in a retrieval system. . transmitted. the link order in this example should be set so that the single header row is assigned to the Right input.Use Entire partitioning on the header input to copy the single header row to all partitions Figure 31: Optimized Partitioning assignment based on business requirements Because we are joining on a single value. the above solution has one serious limitation. transcribed. In order to process a large number of detail records. the single value join column will assign all rows to a single partition. so we will revisit this in the Sorting discussion. consider that the single header row is really a form of reference data. To optimize partitioning. An optimized solution would be to alter the partitioning for the input links to the Join stage: . All rights reserved. it is not always the most efficient solution. resulting in sequential processing. depending on the business requirements.

5. there is a specialized example where the Round Robin collector may be appropriate. Auto is the default collector method. 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. transmitted.5.Information Integration Solutions Center of Excellence If defined in reverse of this order. No part of this publication may be reproduced. 2006 72 of 179 © 2006 IBM Information Integration Solutions. . Like partitioning methods. 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. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially.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. the order of rows in an Auto collector is undefined. 5. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition. or translated into any language in any form by any means without the written permission of IBM. the collector method is defined in the stage Input/Partitioning properties for any stage running sequentially. (See Section 12. All rights reserved. transcribed. 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. 5. For this reason. stored in a retrieval system.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.3: Minimizing Runtime Processes and Resource Requirements). For advanced users. and may vary between job runs on the same Data Set. there is one further detail in this example. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5. the detail rows in the Left input will buffer to disk to prevent a deadlock. However.

5.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. and these should be the same columns. as used to sort the input Data Set in parallel. Parallel Framework Red Book: Data Flow Job Design July 17.6 Collecting Methodology Given the options for collecting data into a sequential stream. an Ordered collector will generate a sequential stream in sort order. In this scenario.5. transcribed. transmitted. This is because Round Robin collector reads from partitions using the same partition order that a Round Robin partitioner assigns rows to parallel partitions. All rights reserved. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. the Sort Merge collector will generate a sequential stream of rows in globally sorted order. 5. as long as the Data Set has not been repartitioned or reduced.3 Ordered Collector An Ordered collector reads all rows from the first partition. in the same order. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. 5. stored in a retrieval system. Row order is undefined for non-key columns. then reads all rows from the next partition until all rows in the Data Set have been collected. or translated into any language in any form by any means without the written permission of IBM. 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. 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.5. through aggregation). No part of this publication may be reproduced. The Sort Merge collector requires one or more key column(s) to be defined.4 Sort Merge Collector If the input Data Set is sorted in parallel. . 2006 73 of 179 © 2006 IBM Information Integration Solutions. Ordered collectors are generally only useful if the input Data Set has been Sorted and Range partitioned on the same key column(s). use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. use Sort Merge collector to produce a single.

Information Integration Solutions Center of Excellence 6 Sorting Traditionally. Sort Aggregator. 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. This is the method the SQL databases use for an ORDER BY clause. sort order is not maintained. as illustrated in the following example. 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. a sort is required after repartitioning. Merge) require pre-sorted groups of related records. 6. 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. transmitted.Sorting: is used to establish group order within each partition. stored in a retrieval system. For example. ordered result set. All rights reserved. In the following example. To restore row order and groupings.Partitioning: is used to gather related records. 2006 74 of 179 © 2006 IBM Information Integration Solutions. This sort can be done in parallel. Other stages (for example. the previous input Data Set is partitioned on LName and FName columns. The order of key columns determines the sequence and groupings in the result set. when data is re-partitioned.1 Partition and Sort Keys Using the parallel Sort within DataStage Enterprise Edition: . based on one or more key column(s) NOTE: By definition. sorting on primary key LName (ascending). Join. Change Apply. in most cases there is no need to globally sort data to produce a single sequence of rows. sorting is most often needed to establish order within specified groups of data. Change Capture. the process of sorting data uses one primary key column and. 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. assigning rows with the same key column values to the same partition . we would see the following results: Parallel Framework Red Book: Data Flow Job Design July 17. Given a 4-node configuration file. . optionally. one or more secondary key column(s) to generate a sequential. Instead. transcribed. Each column is specified with an ascending or descending sort order.

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

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

the Sort Key Mode is most frequently used. No part of this publication may be reproduced. Specifically. the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization. 2006 77 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence - Position of nulls in the result set (for nullable columns) Figure 36: Specifying Link Sort options 6.3.5 Sub-Sorts Within the standalone Sort stage. For this reason. This “subsort” uses significantly less disk space and CPU resource. It is used when resorting a sub-grouping of a previously sorted input Data Set. 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.4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. and can often be performed in memory (depending on the size of the new subsort groups). stored in a retrieval system. disable Stable sort unless needed. NOTE: The Sort Utility option is an artifact of previous releases. - 6. . Always specify “DataStage” Sort Utility. 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. transmitted. This requires some additional overhead in the sort algorithm.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. Parallel Framework Red Book: Data Flow Job Design July 17. which is significantly faster than a “UNIX” sort. instead of performing a complete Sort. or translated into any language in any form by any means without the written permission of IBM. 6. transcribed. All rights reserved. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys.

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

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

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

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

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

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. Note that this document is required.3.3. When multiple files are read by a single Sequential File stage (using multiple files. or by using a File Pattern). as shown below: Parallel Framework Red Book: Data Flow Job Design July 17. making the entire downstream flow run sequentially (unless it is later repartitioned). The format of the Schema File. transcribed. Setting this value to a low number (such as 1) is useful for realtime applications.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. 7. the Sequential File (export operator) stage buffers its writes to optimize performance. No part of this publication may be reproduced. 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. instead of statically through Table Definitions. or translated into any language in any form by any means without the written permission of IBM. SAME will read the entire file into a single partition. transmitted. 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.3. 7. It is also important to remember that this setting will apply to all Sequential File stages in the data flow. 2006 83 of 179 © 2006 IBM Information Integration Solutions.3. Using the Schema File option allows the format of the source file to be specified at runtime. • 7. stored in a retrieval system. the buffers are always flushed to disk. It is important to use ROUNDROBIN partitioning (or other partitioning appropriate to downstream components) to evenly distribute the data in the flow. but there is a small performance penalty associated with increased I/O. 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. All rights reserved. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. each file’s data is read into a separate partition. . When a job completes successfully.5 Sequential File (Export) Buffering By default. 7.7 Reading and Writing Nullable Columns When reading from or writing to Sequential Files or File Sets.

Double-click on the column number in the grid dialog to set this column property. Decimal. Varchar). or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. if the source file has fields with values longer than the maximum Parallel Framework Red Book: Data Flow Job Design July 17.8 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. use the print_field import property. Use caution when specifying this option as it can generate an enormous amount of detail in the job log. Double-click on the column number in the grid dialog to set this column property.9 Reading Bounded-Length VARCHAR Columns Care must be taken when reading delimited. • • 7.3. All rights reserved. the field width and pad string column properties must be set to match the fixed-width of the output column. you must define the null field value and length in the Nullable section of the column property. Integer. transmitted. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 40: Extended Column Metadata (Nullable properties) 7. 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. Decimal. transcribed. All import and export properties are listed in the Import/Export Properties chapter of the Orchestrate Operators Reference. Integer. the field width column property must be set to match the fixed-width of the input column. Varchar). When writing fixed-length files from variable-length fields (eg. To display each field value. • If a field is nullable. By default. . bounded-length Varchar columns (Varchars with the length option set). 2006 84 of 179 © 2006 IBM Information Integration Solutions.3.

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

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

For this reason. . To create a Transformer reject link in Designer. 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.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer. For this reason. by definition.col Note that if an incoming column is only used in an output column mapping. or translated into any language in any form by any means without the written permission of IBM. See Section 3. entries are placed in the Director job log. 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.col) Then “” Else link. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link. For example.Information Integration Solutions Center of Excellence 8 Transformation Languages 8. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17. 8.1. This makes it easy to identify reject conditions (by row counts). 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. 2006 87 of 179 © 2006 IBM Information Integration Solutions.4: Parallel Transformer stages for guidelines on Transformer stage usage. 8. stored in a retrieval system. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. Because the DS/EE Transformer runs in parallel. transcribed. the Transformer will reject the row instead of sending it to the output link(s). if you intend to use a nullable column within a Transformer derivation or output link constraint. All rights reserved.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints.8. undefined. No part of this publication may be reproduced. Always include reject links in a parallel Transformer. transmitted.1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. it is important to minimize the number of transformers. Even if the target column in an output derivation allows nullable results.1. When rows are rejected by a Transformer.

If more precision is required. 8. Within the link constraints dialog box. By default. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7. the incoming column must be type VarChar before it is evaluated in the Transformer. 2006 88 of 179 © 2006 IBM Information Integration Solutions. it is important to make sure the type conversion is done before a row reaches the Transformer. Therefore. apply the abort rule to this output link.1. transmitted. up to a maximum precision of 255 and scale of 125.5: Sequential File (Export) Buffering). not the target.4 -> 1. 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. For example.6 ->1.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment.4 -> 2. and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example. Examples: 1. these internal decimal variables will have a precision and scale of 38 and 10. -1.3. Thus. breaking ties by rounding positive values toward positive infinity and negative values toward negative infinity. the PadString function uses the length of the source type. -1. Examples: 1.1.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage.6 -> -1 floor Rounds towards negative infinity.4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition. When generating a sequence of numbers in parallel.4 -> -2 round_inf Rounds or truncates towards nearest representable value.4 -> -1.1. For example. 8. 1. internal decimal results are rounded to the nearest applicable value.5-> 2. or performing parallel derivations. -1. Since the Transformer will abort the entire job flow immediately.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17. . there are times when internal decimal variables need to be generated in order to perform the evaluation. 8. Create a new output link that will handle rows that match the abort rule. TrimLeadingTrailing(string) works only if string is a VarChar field. stored in a retrieval system. No part of this publication may be reproduced. -1. or committed to database tables. transcribed. By default. Examples: 1. 1). it is possible that valid rows will not have been flushed from Sequential File (export) buffers. the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range. All rights reserved. the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. or translated into any language in any form by any means without the written permission of IBM.

Examples: 1.5. The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value. -1.1. This can be made more efficient by moving the substring calculation into a stage variable. 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.56 ->-1. In this case. then the following test may appear in a number of output columns derivations: IF (DSLINK1. No part of this publication may be reproduced. Similarly.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign. transcribed. transmitted. it is useful to understand what items get evaluated and when. In this case. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. 8.3] = “001”) THEN . the stage variable definition would be: Parallel Framework Red Book: Data Flow Job Design July 17. By doing this.. For example.56 -> 1. .3] is evaluated for each column that uses it.col[1. as they would be evaluated once for every output column that uses them. round or truncate to the scale size.5. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. it can be seen that there are certain constructs that would be inefficient to include in output column derivations. the evaluation of the substring of DSLINK1. unless the derivation is empty For each output link: Evaluate the link constraint.col[1. All rights reserved.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations. 2006 89 of 179 © 2006 IBM Information Integration Solutions. suppose multiple columns in output links want to use the same substring of an input column. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example. From this sequence. the output links are also evaluated in the order in which they are displayed. the substring is evaluated just once for every input row..

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

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

value is the value you wish to represent a null in the output.dataType is its optional data type. The character argument is the character to remove. . . date. All rights reserved. By default. transmitted.Information Integration Solutions Center of Excellence 8. with the following syntax: stringField=string_trim[character. To convert from an in-band null to an out-of-band null. for decimal. No part of this publication may be reproduced. The destField is converted from an Orchestrate out-of-band null to a value of the field’s data type. To convert from an out-of-band null to an in-band null (value) representation within Modify.dataType is its optional data type.2 Modify and String Trim The function string_trim has been added to Modify. these characters are retained when the fixed-length string is then converted back to a variable-length string. . use it if you are also converting types. or translated into any language in any form by any means without the written permission of IBM.2. the syntax is: destField[:dataType] = handle_null (sourceField. and justify defaults to begin. stored in a retrieval system. the syntax is: destField[:dataType] = make_null(sourceField. transcribed.2. NOTE: The DataStage Parallel Job Developers Guide gives incorrect syntax for converting an out-of-band null to an in-band null (value) representation. direction defaults to end. use it if you are also converting types.destField is the destination field’s name. 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 17.1 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. string.value is the value of the source field when it is null. direction. this is NULL. value can be a string. 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. 8. By default. Justify has no affect when the target string has variable length. . For a numeric field value can be a numeric value.value) where: . 2006 92 of 179 © 2006 IBM Information Integration Solutions.destField is the destination field’s name.sourceField is the source field’s name.value) where: . . and timestamp fields. The value of the direction and justify arguments can be either begin or end. .sourceField is the source field’s name . time.

. begin](color) Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved.Information Integration Solutions Center of Excellence name:string = string_trim[NULL. end. 2006 93 of 179 © 2006 IBM Information Integration Solutions. begin](name) The following example removes all trailing Z characters from color. transcribed. 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. transmitted. and left-justifies the resulting hue fixed-length string: hue:string[10] = string_trim[‘Z’.

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

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 output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double). 9. set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. aggregate on generated key column) there is no need to sort or hash-partition the input data with only one key column value aggregate Sequentially on the generated column - - Note that in this example use two Aggregators are used to prevent the sequential aggregation from disrupting upstream processing. The Sort Aggregation Method should be used when the number of key values is unknown or very large. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property.3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings. which maintains the results of each keycolumn value/aggregation pair in memory. Unlike the Hash Aggregator. 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. Because each key value/aggregation requires approximately 2K of memory.3. To perform a total aggregation. 2006 95 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.3. Parallel Framework Red Book: Data Flow Job Design July 17.3 The Aggregator Stage 9. To aggregate in decimal precision. the Hash Aggregator should only be used when the number of distinct key values is small and finite. transcribed. All rights reserved. An exception to this is financial calculations which should be done in decimal to preserve appropriate precision. stored in a retrieval system.1 Aggregation Method By default. the default Aggregation Method is set to Hash. or translated into any language in any form by any means without the written permission of IBM. the Sort Aggregator requires presorted data.Information Integration Solutions Center of Excellence 9. transmitted.3. .2 Aggregation Data Type By default. 9. but only maintains the calculations for the current group in memory.

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

and version. Because plug-in stage types cannot read in parallel. This allows Enterprise Edition to match return columns by name. transmitted. 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). not position in the stage Table Definitions.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. The benefit of ODBC Enterprise stage comes from the large number of included and third party ODBC drivers to enable connectivity to all major database platforms. not name. From a design perspective. so Table Definitions must match the order of columns in a query.1.1. the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server. 10. Unlike the database-specific parallel stages. stored in a retrieval system. they should only be used when it is not possible to use a native parallel stage. . All rights reserved. 10.1. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect. For example.1. specific guidelines of when to use various stage types are provided in the database-specific topics in this section. for maximum parallel performance. native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform. and cannot span multiple servers in a clustered or Grid configuration. operating system. Because of their tight integration with database technologies. or translated into any language in any form by any means without the written permission of IBM. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. attributes). care must be taken to assign the correct data types in the job design. No part of this publication may be reproduced.1. and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied. the ODBC Enterprise stage cannot interface with database-specific parallel load technologies.1. 2006 97 of 179 © 2006 IBM Information Integration Solutions. the native parallel stages often have more stringent connectivity requirements than plug-in stages. However. transcribed. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform.Information Integration Solutions Center of Excellence 10. Furthermore. plug-in database stages match columns by order. the DataDirect ODBC Driver Manager is also included. On UNIX. scalability. Parallel Framework Red Book: Data Flow Job Design July 17.2 ODBC Enterprise stage In general.1 Native Parallel database stages In general. Native parallel stages always pre-query the database for actual runtime metadata (column names. Because there are exceptions to this rule (especially with Teradata). types.

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

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

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

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

it can still be beneficial to use database filters (WHERE clause) if appropriate. transcribed. All rights reserved. For maximum scalability and parallel performance. especially when the join columns are indexed. Oracle PL/SQL) on a per-row basis within a high-volume data flow. No part of this publication may be reproduced. it is best to implement business rules using native parallel DataStage components.Information Integration Solutions Center of Excellence • Use a SQL Join to combine data from tables with a small number of rows in the same database instance. 2006 102 of 179 © 2006 IBM Information Integration Solutions. A join that reduces the result set significantly is also often appropriate to do in the database. Avoid the use of database stored procedures (for example. stored in a retrieval system. or when the source includes a large number of database tables. In this scenario. transmitted. • • 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. the efficiency of the Enterprise Edition Sort and Join stages can be significantly faster than an equivalent SQL query. . When combining data from very large tables.

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

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

or translated into any language in any form by any means without the written permission of IBM. 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. Enter the password for connecting to DB2. transmitted. • Server. 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.Information Integration Solutions Center of Excellence Figure 45: DB2/UDB Enterprise stage properties For connection to a remote DB2/UDB instance. Optionally set this to the instance name of the DB2 server. it is possible to connect to more than one DB2 instance within a single job.cfg file can July 17. • Client Alias DB Name.2. Otherwise use the DB2 environment variable. . Since a db2nodes. If you set this property.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. you will need to set the following properties on the DB2/UDB Enterprise stage in your parallel job: • Client Instance Name. 10. No part of this publication may be reproduced. Enter the user name for connecting to DB2. Set this to the DB2 client’s alias database name for the remote DB2 server database. Single stream . • Password. DataStage assumes you require remote connection. to identify the instance name of the DB2 server. Once this variable is set. DB2INSTANCE. All rights reserved. Your job must meet one of the following configurations (note: the use of the word “stream” refers to a contiguous flow of one stage to another within a single job): 1. This is required only if the client’s alias is different from the actual name of the remote server database. it will try to use it for each of the connections in the 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. Optionally set this to the remote server database name. Set this to the DB2 client instance name. 2006 105 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. stored in a retrieval system.3 Configuring DB2 Multiple Instances in One DataStage Job Although it is not officially supported. Otherwise use the environment variables $APT_DBNAME or $APT_DB2DBDFT to identify the database. 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. transcribed. • User. 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. • Database.

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

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

10. In this configuration. In these instances. reading from and writing to DB2 in parallel (where appropriate). transcribed. communicating directly with each database node. forcing the DB2 Enterprise stages on each ETL server to communicate directly with their local DB2 coordinator. During the load operation. When writing to a DB2 database in parallel. and writing to a DB2 Enterprise Server Edition database with Database Partitioning Feature (DPF) • As a native. The choice between these methods depends on required performance.2. and recoverability. parallel component. transmitted. the DB2/UDB Enterprise stage is designed for maximum performance and scalability against very large partitioned DB2 UNIX databases. The DB2 load operator performs a non-recoverable load.7 DB2/UDB Enterprise stage options The DB2/UDB Enterprise (native parallel) stage should be used for reading from.8 Performance Notes In some cases. and the target table(s) may be accessed by other users. DB2 Enterprise stage should not include the Client Instance Name property.Information Integration Solutions Center of Excellence 10.2. 10. No part of this publication may be reproduced. the DB2/UDB Enterprise stage offers the choice of SQL (insert / update / upsert / delete) or fast DB2 loader methods. and the availability of new rows to other applications. All rights reserved.2. b) The DB2 Load method requires that the DataStage user running the job have DBADM privilege on the target DB2 database. database log usage. if the load operation is terminated before it is completed. 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. performing lookups against. All operations are logged to the DB2 database log. stored in a retrieval system. a) The Write Method (and corresponding insert / update / upsert / delete) communicates directly with the DB2 database nodes to execute instructions in parallel. when using user-defined SQL without partitioning against large volumes of DB2 data.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. 2006 108 of 179 © 2006 IBM Information Integration Solutions. That is. All activity in the z/OS environment always goes through the DB2 coordinator node so parallelism differs slightly depending on how DB2 is accessed. • DB2/UDB Enterprise stage is tightly integrated with the DB2 RDBMS. Time and row-based commit intervals determine the transaction size. . and using the same data partitioning as the referenced DB2 tables. 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). the DB2 Load DataStage job must be re. the overhead of routing information through a remote DB2 coordinator may be significant.run in Truncate mode to clear the load pending state. 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. In this scenario. the contents of the table are unusable and the tablespace is left in a load pending state.

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

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

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

. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. transcribed.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. All rights reserved. stored in a retrieval system. transmitted. 2006 112 of 179 © 2006 IBM Information Integration Solutions.

3. decimal[p. time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month.s] where p is the precision and s is the scale The maximum precision is 32. the result is a date field.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: .3.r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date. as shown in the following table. or translated into any language in any form by any means without the written permission of IBM.columns of the database row correspond to columns of a DS/EE record . stored in a retrieval system.3 Informix Database Guidelines 10.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n. DECIMAL[p.Information Integration Solutions Center of Excellence 10. Informix Data Type CHAR(n) CHARACTER VARYING(n. If the DATETIME starts with a year component.both DS/EE and Informix 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. 2006 113 of 179 © 2006 IBM Information Integration Solutions.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. . transcribed. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. All rights reserved.2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. the result is a timestamp field. The Informix Enterprise stage converts Informix data types to Enterprise Edition data types.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 .r) NVARCHAR(n. A decimal with floating scale is converted to dfloat dfloat dfloat int32 decimal string[n] string[max=n] sfloat int32 sfloat int16 string[max=n] IMPORTANT: Informix data types that are not listed in the above table cannot be used in the Informix Enterprise stage. No part of this publication may be reproduced. the result is a time field.rows of the database result set correspond to records of a DS/EE Data Set .

stored in a retrieval system.2 ODBC Enterprise stage Data Type Mapping ODBC data sources are not supported by the orcdbutil utility. In that case. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.4. All rights reserved. .s] string[36] int32 int8 [0 or 1] decimal[p.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. 2006 114 of 179 © 2006 IBM Information Integration Solutions.1 ODBC Enterprise Stage Column Names For each ODBC Enterprise stage: .columns of the database row correspond to columns of a DS/EE record .s] decimal[p. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. If you attempt to read a record larger than 32K.names are translated exactly except when the external data source column name contains a character that DataStage does not support. It is important to verify the correct ODBC to Enterprise Edition data mapping.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 .Information Integration Solutions Center of Excellence 10. transmitted.rows of the database result set correspond to records of a DS/EE Data Set . 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. 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.4 ODBC Enterprise Guidelines 10. Enterprise Edition will return an error and abort your job.both DS/EE and ODBC support null values.4. two underscore characters replace the unsupported character .s] where p is the precision and s is the scale decimal[p.

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

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

If you attempt to read a record larger than 32K. All rights reserved.names are translated exactly except when the Oracle source column name contains a character that DataStage does not support.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p.5. transmitted.5 Oracle Database Guidelines 10. The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.5.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. Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p.5. and will generate an error at runtime 10. stored in a retrieval system. Examples include: . Setting the partition table option to the specified table will enable parallel extracts from an Oracle source.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 . 2006 117 of 179 © 2006 IBM Information Integration Solutions.s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38. This is particularly important for Oracle databases. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage. the Oracle Enterprise stage reads sequentially from its source table or query.columns of the database row correspond to columns of a DS/EE record .2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.rows of the database result set correspond to records of a DS/EE Data Set .Information Integration Solutions Center of Excellence 10. two underscore characters replace the unsupported character .3 Reading from Oracle in Parallel By default. It is important to note that certain types of queries cannot run in parallel. transcribed. .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.both DS/EE and Oracle support null values.1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . which are not heavily typed. In that case. Enterprise Edition will return an error and abort your job.

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

Information Integration Solutions Center of Excellence

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

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

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

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

Information Integration Solutions Center of Excellence

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

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

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

Parallel Read
Yes No

Teradata Utility Limit
applies none

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

Parallel Write
Yes

Teradata Utility Limit
applies

Teradata MultiLoad (MultiLoad utility)

Plug-In

No

applies

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

120 of 179

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

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

Teradata API

Plug-In

Yes

none

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

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

Parallel Framework Red Book: Data Flow Job Design

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

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

A-Z (case insensitive). where x is 0-9.7. $. Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. transcribed. All rights reserved. No part of this publication may be reproduced. the tdpid must be in the form TDPx. The first three characters must be TDP. you must supply the client with the Teradata Director Program (TDP) identifier. the tdpid is the host name of the Teradata server.Information Integration Solutions Center of Excellence 10. That leaves 39 possible TDP names and is different than the convention used for non-channel attached systems. also known as the tdpid. On a network-attached system. or @. On MVS. stored in a retrieval system. To connect to a Teradata server.9 Teradata on USS On the USS platform the Teradata Enterprise Stage uses CLIv2 for channel-attached systems (OS/390 and z/OS). transmitted. . #. 2006 123 of 179 © 2006 IBM Information Integration Solutions.

No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 11 Troubleshooting and Monitoring 11.2 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. the job’s partitioning logic and parallel design concepts (especially within Transformer stages) should be examined. This will ensure that the jobs have been designed with proper partitioning logic. or translated into any language in any form by any means without the written permission of IBM. lowering this number will limit the number of processes that are simultaneously created when a job is started. stored in a retrieval system. including CPU utilization and elapsed processing time. If the job results are correct with a single-node configuration file. 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. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. . This can be useful when determining if the actual runtime schema matches the expected job design table definitions. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. $DS_PX_DEBUG 1 $APT_PM_STARTUP_CONCURRENCY 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. generated osh.1 Warning on Single-Node Configuration Files Because the DS/EE configuration file can be changed at runtime. Set this environment variable to capture copies of the job score. Prints detailed information in the job log for each operator. All rights reserved. this can identify which part of a job flow is impacting overall performance. Forces an internal buffer operator to be placed between every operator. Normally. it is important that all jobs be tested with a configuration file that has more than one node in its default node pool. transcribed. transmitted. Using the job monitor performance statistics. When trying to start very large jobs on heavilyloaded servers. July 17. 11. Prints detailed information in the job log for each operator when allocating additional heap memory. the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance. This environment variable should not normally need to be set. and internal Enterprise Edition log messages in a directory corresponding to the job name. and incorrect with a multi-node configuration file.

All fatal and warning messages should be addressed before attempting to debug. transcribed.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide. stored in a retrieval system. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG . This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server.    NOTE: For large jobs. When a fatal error occurs. 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The general process for debugging a job is:  Check the Director job log for warnings. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. tune. 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. . the log entry is sometimes preceded by a warning condition. generated osh. To capture the full schema output in these cases. and internal Enterprise Edition log messages in a directory corresponding to the job name. July 17.5. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader.  Enable the Job Monitoring Environment Variables detailed in Section 2. look at row counts. Set the environment variable $DS_PX_DEBUG to capture copies of the job score. In some instances. These may indicate an underlying logic problem or unexpected data type conversion. it may not be possible to remove all warning messages generated by the EE engine. 11. 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. The default is 30 seconds.Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. or promote a job from development into test or production. and to manage source or target Parallel Data Sets. All rights reserved. or translated into any language in any form by any means without the written permission of IBM.  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled).3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. But all warnings should be examined and understood. transmitted.

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

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

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

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

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

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

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

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

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

cannot be greater than 1048576. Cannot be less than 8192. 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. Default is 1048576.4. stored in a retrieval system. Here is an example job fragment: Figure 55: Fork-Join example Parallel Framework Red Book: Data Flow Job Design July 17. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players.2 Deadlock Prevention Buffering The other type of buffering. 2006 135 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records. transmitted. APT_LATENCY_COEFFICIENT. 12. All rights reserved. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. . “Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job. transcribed. No part of this publication may be reproduced.

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

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

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

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

5: Job Parameters. transmitted. along with cross-references for more detail. user. Parallel Framework Red Book: Data Flow Job Design July 17. This can also be used for integration with source code control systems. No part of this publication may be reproduced.  Create a standard directory structure outside of the DataStage project directory for source and target files.  A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN.3: Documentation and Annotation. See Section 2. as well as Annotation fields. Standards It is important to establish and follow consistent standards in:  Directory structures for install and application support directories. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. database login settings. 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.2: Naming Conventions. transcribed. database login properties…)  environment variables and their default settings (as outlined in Section 2. stage names. 2. password. source and target file paths.4: Working with Source Code Control Systems. intermediate work files. DataStage Template jobs should be created with:  standard parameters (for example. and so forth. file names. It is the DataStage developer’s responsibility to make personal backups of their work on their local workstation. or translated into any language in any form by any means without the written permission of IBM. .  Where possible. especially for DataStage Project categories.1 Environment Variable Settings)  annotation blocks Job Parameters should always be used for file paths. An example directory naming structure is given in Section 2.1:Directory Structures. using the Manager DSX export capability.  Naming conventions. All rights reserved. 1. etc) and directories where files are stored. these standard parameters and settings should be made part of a Designer Job Template. Development Guidelines Modular development techniques should be used to maximize re-use of DataStage jobs and components. To ease re-use. and links.Information Integration Solutions Center of Excellence Appendix A: Standard Practices Summary This Appendix summarizes Standard Practices recommendations outlined in this document.5. All DataStage jobs should be documented with Short Description field. The scope of a parameter is discussed further in Section 3. An example DataStage naming structure is given in Section 2. stored in a retrieval system. 2006 140 of 179 © 2006 IBM Information Integration Solutions. create re-usable components such as parallel shared containers to encapsulate frequently-used logic. See Section 2.

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

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

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. 2006 143 of 179 © 2006 IBM Information Integration Solutions. use Sort Merge collector to produce a single. See Section 9. as long as the Data Set has not been repartitioned or reduced. No part of this publication may be reproduced.  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. use the Join or Merge stage. All rights reserved.Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel. 7. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned.1. Join vs. transmitted. transcribed. stored in a retrieval system. 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. 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. Sorting Using the rules and behavior outlined in Section 6: Sorting.  Stage-Specific Guidelines As discussed in Section 8. precautions must be taken when using expressions or derivations on nullable columns within the parallel Transformer: o Always convert nullable columns to in-band values before using them in an expression or derivation.1: Lookup vs. If the Data Sets are larger than available memory resources. . Merge.1: Transformer NULL Handling and Reject Link. The Lookup stage is most appropriate when reference data is small enough to fit into available memory. Create Cluster Key Change Column. Create Key Change Column. o Always place a reject link on a parallel Transformer to capture / audit possible rejects. sequential ordered result set use a parallel Sort and a Sort Merge collector 8.

Check the Director log for warnings. A Sort method Aggregator should be used when the number of distinct key values is large or unknown. stored in a retrieval system. transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. (See Section 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. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. or Informix databases. use a SQL where clause to limit the number of rows sent to a DataStage job. 10. Use Hash method Aggregators only when the number of distinct key column values is small. use the native parallel database stages for maximum performance and scalability. 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. No part of this publication may be reproduced. 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. which may indicate an underlying problem or data type conversion issue. or when exception processing.  Database Stage Guidelines Where possible. For maximum scalability and parallel performance. 2006 144 of 179 © 2006 IBM Information Integration Solutions.1. When using Oracle. All rights reserved. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. 9. transcribed. or translated into any language in any form by any means without the written permission of IBM. If possible. it is best to implement business rules natively using DataStage parallel components. as discussed in Section 10. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow.   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.2Capturing Unmatched Records from a Join). . the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records. Even if the source data is not nullable. DB2.

transcribed. 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. . error and warning messages from a running DS/EE job.Information Integration Solutions Center of Excellence    The environment variable $DS_PX_DEBUG can be used to capture all generated OSH. 2006 145 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.2 Understanding the Parallel Job Score. stored in a retrieval system. Enable $APT_DUMP_SCORE by default. Set $DS_PX_DEBUG if the schema record is too large to capture in a Director log entry. All rights reserved. No part of this publication may be reproduced. and examine the job score by following the guidelines outlined in Section 12.4.

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

.Information Integration Solutions Center of Excellence XML Output XML Transformer Restructure Stages Column Export Column Import XMLo XMLt CExp CImp Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. stored in a retrieval system. transcribed. All rights reserved. 2006 147 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.

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

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

not SAME) Parallel producer to Sequential consumer No producer or no consumer (typically. 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. The memory used by deadlock-prevention BufferOps can be calculated based on the number of inserted BufferOps. 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. transcribed. collector type with the second. 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. The partition type is associated with the first term. . All rights reserved. for persistent Data Sets) July 17. Within the Data Set definition. For example. Producers and consumers may be either persistent (on disk) Data Sets or parallel operators. followed by the downstream consumer. corresponding to the lower section of the job score. the first Data Set is identified as “ds0”. In the above example. stored in a retrieval system. transmitted. and the next “ds1”. while operators are identified by their operator number and name. 2006 150 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. 3. Data Sets are identified in the first section of the parallel job score. the upstream producer is identified first. with 1 degree of parallelism [1p]. 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. with each Data Set identified by its number (starting at zero). Persistent Data Sets are identified by their Data Set name. operator zero (op0) is running sequentially. in the example on the right. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Note that the number of virtual Data Sets and the degree of parallelism determine the amount of memory used by the inter-operator transport buffers. Operator 1 (op1) is running in parallel with 4 degrees of parallelism [4p]. followed by a notation to indicate the type of partitioning or collecting (if any).

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

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

or translated into any language in any form by any means without the written permission of IBM. the method APT_Record::estimateFinalOutputSize() can give you an estimate for a given record schema. Parallel Framework Red Book: Data Flow Job Design July 17. 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: . 2006 153 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. if you have a transfer that transfers all fields from input to output.add (# nullable fields)/8 for null indicators . All rights reserved. No part of this publication may be reproduced. . 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)/2.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries. transcribed. transmitted.one byte per column for field alignment (worst case is 3.Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. As can APT_Transfer::getTransferBufferSize().

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

this will significantly improve file pattern reads.000 bytes. If it is not found the importer looks ahead 4*500=2000 (1500 more) bytes. $DB2DBDFT is used to find the database name. Allows DataStage plug-in stages to handle DB2 databases which use the special characters # and $ in column names. Specifies the number of records to insert between commits. but this can be set as low as 2 bytes. For disk configurations with multiple controllers and disk. . This variable controls the upper bound which is by default 100.000 bytes. Sequential File (import) will read ahead 500 bytes to get the next delimiter. stored in a retrieval system. 2006 155 of 179 © 2006 IBM Information Integration Solutions. use this variable instead of APT_DELIMITED_READ_SIZE. transcribed.db2profile. The default is 500 bytes. When more than 500 bytes read-ahead is desired. and so on (4X) up to 100. Environment Variable $INSTHOME DB2 Environment Variables Setting [path] [path] [database] Description Specifies the DB2 install directory. This setting should be set to a lower value when reading from streaming inputs (for example. This variable is usually set in a user’s environment from . 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. No part of this publication may be reproduced. transmitted. Used as a backup for specifying the DB2 installation directory (if $INSTHOME is undefined). or translated into any language in any form by any means without the written permission of IBM. The default value is 2000 per partition. By default. $APT_DB2INSTANCE_HOME $APT_DBNAME $APT_RDBMS_COMMIT_ROWS Can also be specified with the “Row Commit Interval” stage input property. All rights reserved.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. Specifies the name of the DB2 database for DB2/UDB Enterprise stages if the “Use Database Environment Variable” option is True. $DS_ENABLE_RESERVED_CHAR_CONVE RT [rows] 1 Parallel Framework Red Book: Data Flow Job Design July 17. socket or FIFO) to avoid blocking. If $APT_DBNAME is not defined. $APT_MAX_DELIMITED_READ_SIZE [bytes] $APT_IMPORT_PATTERN_USES_FILESET [set] 3.

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

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.

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

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

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

as well as the average transaction amount per-item. and append these values to the original data. per store for all stores in the nation. 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. 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. No part of this publication may be reproduced. . 2006 163 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. Here is our source data: Data Set 1: 32 Rows Store Location 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Item ID 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 Transaction Date 2004-01-01 2004-01-02 2004-01-01 2004-01-03 2004-01-04 2004-01-02 2004-01-04 2004-01-03 2004-01-04 2004-01-03 2004-01-01 2004-01-04 2004-01-02 2004-01-02 2004-01-01 2004-01-03 2004-01-03 2004-01-01 2004-01-02 2004-01-04 2004-01-02 2004-01-03 2004-01-01 2004-01-04 2004-01-04 2004-01-01 2004-01-03 2004-01-02 2004-01-04 2004-01-02 2004-01-01 2004-01-03 Transaction Amt 1 2 3 5 5 54 7 8 2 3 45 65 7 85 9 98 23 3 32 45 54 56 7 8 23 45 534 6 65 7 78 87 The screen capture below shows how to implement the business logic in an efficient manner. transcribed. transmitted. They wish to determine the weighted average transaction amount peritem nation-wide. thereby increasing the value of this exercise. stored in a retrieval system.Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets 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.

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

it guarantees the output order to be sorted since the result of aggregation can be released for downstream processing as soon as the key change is detected. not Hash.5 39. you will need to set the Aggregator’s “Method” to Sort.5 50.5 6. transmitted.5 35. in return. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’. stored in a retrieval system.5 310.5 35.5 310.5 Store Average Item Transaction Amt 28 28 43 43 6. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 180. however.5 46 46 6.5 23 23 23 23 26.5 3.5 180. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 6.5 180.5 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.25 26.5 3. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation. it does not guarantee output order. . It does this by keeping running totals in memory for the aggregation for each output group. or translated into any language in any form by any means without the written permission of IBM.25 26.25 16.5 35.25 26. 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. transcribed.5 39.25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17. The hash method only requires the input data to be hashed.5 3.5 16. Therefore. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence NOTE: For this scenario. 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. All rights reserved. The sort method requires the input data to be hashed and sorted.5 180.5 50.

25 44.25 38.25 44.75 38.75 38.e.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 16.25 44. No part of this publication may be reproduced.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 26. 6. more efficientv. transcribed.5 33. i. and 3.75 38.5 27 27 61. All rights reserved. and 1. A much better answer is: 1.: record adjacency assumes partition concurrency8. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition.5 61. An examination of the job above would suggest: 6. transmitted. 2006 166 of 179 © 2006 IBM Information Integration Solutions.25 44. and 6. 3. or translated into any language in any form by any means without the written permission of IBM. . stored in a retrieval system. 3. This is certainly an improvement on the previous answer.5 33. Here’s a screen shot of this. as only partition concurrency is affected by sequential execution.25 16. Parallel Framework Red Book: Data Flow Job Design July 17. appended to the end of this document for masochistsiv) might suggest: 4. even sequential job execution does not alter the answer for the sort-related questions. A deeper examination (of the score dump.

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

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

5 16.5 33.5 3.75 38.25 16. stored in a retrieval system.25 3.5 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.5 26.75 38. as it is much more expensive than non-stable sorts.5 27 27 61.25 44. useful for preserving previous sort orderings).75 Store Average Item Transaction Amt 2 2 6 6 5 5 26. transcribed. . or translated into any language in any form by any means without the written permission of IBM.25 44. 2006 169 of 179 © 2006 IBM Information Integration Solutions.5 61. No part of this publication may be reproduced. This sort will look exactly like the other sort stage: Remember to disable the ‘Stable Sort’ option if you do not need it (it will try to maintain row order except as needed to perform the sort.75 38.25 44.5 16.Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt. All rights reserved.5 33. Output from above solution: Data Set 3: PeekFinalOutput.25 44. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.5 3.25 38.25 16. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17. and it is enabled by default. transmitted.

25 180.5 35. we had to sort (on disk) 300. transmitted.5 39. but there is a critical difference.5 Store Average Item Transaction 28 28 43 43 6. 000.25 26.5 This solution produces the same result but is achieved with only one complete sort. With the initial solution.25 26.5 46 46 6.5 23 23 180.000. No part of this publication may be reproduced.5 35. It looks very similar to the first solution. and only hashes. There is an even more efficient solution. . Imagine a job with 100 million records as the input.000. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.25 26.5 6.000 records. 2006 170 of 179 © 2006 IBM Information Integration Solutions.5 180. All rights reserved.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume.000.5 50. stored in a retrieval system. a single partitioner. 100. a savings of 400.5 35.5 39.5 180. and two sub-sorts—a much more efficient solution for large data volumes. The second solution only sorts (on disk) 100. or translated into any language in any form by any means without the written permission of IBM.000.5 310.5 6.5 23 23 26.5 50.000 records in addition to hashing 300.000 records.000 records.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 35.5 310. That is a LOT of saved processing power.

25 44. transmitted.5 3. Looks a lot like solution 1. transcribed. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator.5 26.5 27 27 July 17. Data Set 4: PeekFinalOutput.25 16. Also.25 16. Comparing the efficiency of this solution with that of number two.5 16.5 3. All rights reserved. automatic sort insertion. 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. but NOT to use it for hashing. stored in a retrieval system.25 16. No part of this publication may be reproduced.5 3. This is a potentially huge savings on large data volumes (remember the previous example). we saved a sub-sort on 100 million records .Information Integration Solutions Center of Excellence NOTE: In this job.a significant savings. or translated into any language in any form by any means without the written permission of IBM. However it will not create additional processes to handle the records and re-order them. Here is the output from this version of the job.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26. .25 44. and automatic partition insertion must be turned off vii. The difference is on DistributeCopiesOfSourceData: Here. for the same reasons. we have chosen to use the StoreLocation column as a part of our sorting key. except w/o the sort on the output of JoinSourceToAggregator. the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed.

5 33.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. transmitted.5 39.25 44.5 39. .25 26.75 38. or translated into any language in any form by any means without the written permission of IBM.75 61.5 180.5 35.5 46 46 6.5 50.75 38.5 44 44 PeekFinalOutput. stored in a retrieval system.75 38.5 23 23 23 23 26.5 33.25 180.25 26. sort.25 26.5 6.5 310. 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. in addition to the heavy penalty paid in disk I/O for using a full sort.5 6.5 Store Average Item Transaction Amt 28 28 43 43 6. All rights reserved. transcribed. No part of this publication may be reproduced.5 180.25 38.5 35.5 50. by definition.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.5 310.5 180.5 35. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35. Here is a screen shot of a sort running on 40 million records: Parallel Framework Red Book: Data Flow Job Design July 17. 2006 172 of 179 © 2006 IBM Information Integration Solutions.

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

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

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

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

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

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

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

Sign up to vote on this title
UsefulNot useful