Informatica PowerExchange (Version 10.

0)

PowerExchange Interfaces for

PowerCenter
Informatica PowerExchange PowerExchange Interfaces for PowerCenter

Version 10.0
November 2015

Copyright (c) 1993-2015 Informatica LLC. All rights reserved.

This software and documentation contain proprietary information of Informatica LLC and are provided under a license agreement containing restrictions on use and
disclosure and are also protected by copyright law. Reverse engineering of the software is prohibited. No part of this document may be reproduced or transmitted in any
form, by any means (electronic, photocopying, recording or otherwise) without prior consent of Informatica LLC. This Software may be protected by U.S. and/or
international Patents and other Patents Pending.

Use, duplication, or disclosure of the Software by the U.S. Government is subject to the restrictions set forth in the applicable software license agreement and as
provided in DFARS 227.7202-1(a) and 227.7702-3(a) (1995), DFARS 252.227-7013©(1)(ii) (OCT 1988), FAR 12.212(a) (1995), FAR 52.227-19, or FAR 52.227-14
(ALT III), as applicable.

The information in this product or documentation is subject to change without notice. If you find any problems in this product or documentation, please report them to us
in writing.
Informatica, Informatica Platform, Informatica Data Services, PowerCenter, PowerCenterRT, PowerCenter Connect, PowerCenter Data Analyzer, PowerExchange,
PowerMart, Metadata Manager, Informatica Data Quality, Informatica Data Explorer, Informatica B2B Data Transformation, Informatica B2B Data Exchange Informatica
On Demand, Informatica Identity Resolution, Informatica Application Information Lifecycle Management, Informatica Complex Event Processing, Ultra Messaging and
Informatica Master Data Management are trademarks or registered trademarks of Informatica LLC in the United States and in jurisdictions throughout the world. All
other company and product names may be trade names or trademarks of their respective owners.

Portions of this software and/or documentation are subject to copyright held by third parties, including without limitation: Copyright DataDirect Technologies. All rights
reserved. Copyright © Sun Microsystems. All rights reserved. Copyright © RSA Security Inc. All Rights Reserved. Copyright © Ordinal Technology Corp. All rights
reserved.Copyright © Aandacht c.v. All rights reserved. Copyright Genivia, Inc. All rights reserved. Copyright Isomorphic Software. All rights reserved. Copyright © Meta
Integration Technology, Inc. All rights reserved. Copyright © Intalio. All rights reserved. Copyright © Oracle. All rights reserved. Copyright © Adobe Systems
Incorporated. All rights reserved. Copyright © DataArt, Inc. All rights reserved. Copyright © ComponentSource. All rights reserved. Copyright © Microsoft Corporation. All
rights reserved. Copyright © Rogue Wave Software, Inc. All rights reserved. Copyright © Teradata Corporation. All rights reserved. Copyright © Yahoo! Inc. All rights
reserved. Copyright © Glyph & Cog, LLC. All rights reserved. Copyright © Thinkmap, Inc. All rights reserved. Copyright © Clearpace Software Limited. All rights
reserved. Copyright © Information Builders, Inc. All rights reserved. Copyright © OSS Nokalva, Inc. All rights reserved. Copyright Edifecs, Inc. All rights reserved.
Copyright Cleo Communications, Inc. All rights reserved. Copyright © International Organization for Standardization 1986. All rights reserved. Copyright © ej-
technologies GmbH. All rights reserved. Copyright © Jaspersoft Corporation. All rights reserved. Copyright © International Business Machines Corporation. All rights
reserved. Copyright © yWorks GmbH. All rights reserved. Copyright © Lucent Technologies. All rights reserved. Copyright (c) University of Toronto. All rights reserved.
Copyright © Daniel Veillard. All rights reserved. Copyright © Unicode, Inc. Copyright IBM Corp. All rights reserved. Copyright © MicroQuill Software Publishing, Inc. All
rights reserved. Copyright © PassMark Software Pty Ltd. All rights reserved. Copyright © LogiXML, Inc. All rights reserved. Copyright © 2003-2010 Lorenzi Davide, All
rights reserved. Copyright © Red Hat, Inc. All rights reserved. Copyright © The Board of Trustees of the Leland Stanford Junior University. All rights reserved. Copyright
© EMC Corporation. All rights reserved. Copyright © Flexera Software. All rights reserved. Copyright © Jinfonet Software. All rights reserved. Copyright © Apple Inc. All
rights reserved. Copyright © Telerik Inc. All rights reserved. Copyright © BEA Systems. All rights reserved. Copyright © PDFlib GmbH. All rights reserved. Copyright ©
Orientation in Objects GmbH. All rights reserved. Copyright © Tanuki Software, Ltd. All rights reserved. Copyright © Ricebridge. All rights reserved. Copyright © Sencha,
Inc. All rights reserved. Copyright © Scalable Systems, Inc. All rights reserved. Copyright © jQWidgets. All rights reserved. Copyright © Tableau Software, Inc. All rights
reserved. Copyright© MaxMind, Inc. All Rights Reserved. Copyright © TMate Software s.r.o. All rights reserved. Copyright © MapR Technologies Inc. All rights reserved.
Copyright © Amazon Corporate LLC. All rights reserved. Copyright © Highsoft. All rights reserved. Copyright © Python Software Foundation. All rights reserved.
Copyright © BeOpen.com. All rights reserved. Copyright © CNRI. All rights reserved.

This product includes software developed by the Apache Software Foundation (http://www.apache.org/), and/or other software which is licensed under various versions
of the Apache License (the "License"). You may obtain a copy of these Licenses at http://www.apache.org/licenses/. Unless required by applicable law or agreed to in
writing, software distributed under these Licenses is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied. See the Licenses for the specific language governing permissions and limitations under the Licenses.

This product includes software which was developed by Mozilla (http://www.mozilla.org/), software copyright The JBoss Group, LLC, all rights reserved; software
copyright © 1999-2006 by Bruno Lowagie and Paulo Soares and other software which is licensed under various versions of the GNU Lesser General Public License
Agreement, which may be found at http:// www.gnu.org/licenses/lgpl.html. The materials are provided free of charge by Informatica, "as-is", without warranty of any
kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose.

The product includes ACE(TM) and TAO(TM) software copyrighted by Douglas C. Schmidt and his research group at Washington University, University of California,
Irvine, and Vanderbilt University, Copyright (©) 1993-2006, all rights reserved.

This product includes software developed by the OpenSSL Project for use in the OpenSSL Toolkit (copyright The OpenSSL Project. All Rights Reserved) and
redistribution of this software is subject to terms available at http://www.openssl.org and http://www.openssl.org/source/license.html.

This product includes Curl software which is Copyright 1996-2013, Daniel Stenberg, <[email protected]>. All Rights Reserved. Permissions and limitations regarding this
software are subject to terms available at http://curl.haxx.se/docs/copyright.html. Permission to use, copy, modify, and distribute this software for any purpose with or
without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.

The product includes software copyright 2001-2005 (©) MetaStuff, Ltd. All Rights Reserved. Permissions and limitations regarding this software are subject to terms
available at http://www.dom4j.org/ license.html.

The product includes software copyright © 2004-2007, The Dojo Foundation. All Rights Reserved. Permissions and limitations regarding this software are subject to
terms available at http://dojotoolkit.org/license.

This product includes ICU software which is copyright International Business Machines Corporation and others. All rights reserved. Permissions and limitations
regarding this software are subject to terms available at http://source.icu-project.org/repos/icu/icu/trunk/license.html.

This product includes software copyright © 1996-2006 Per Bothner. All rights reserved. Your right to use such materials is set forth in the license which may be found at
http:// www.gnu.org/software/ kawa/Software-License.html.

This product includes OSSP UUID software which is Copyright © 2002 Ralf S. Engelschall, Copyright © 2002 The OSSP Project Copyright © 2002 Cable & Wireless
Deutschland. Permissions and limitations regarding this software are subject to terms available at http://www.opensource.org/licenses/mit-license.php.

This product includes software developed by Boost (http://www.boost.org/) or under the Boost software license. Permissions and limitations regarding this software are
subject to terms available at http:/ /www.boost.org/LICENSE_1_0.txt.

This product includes software copyright © 1997-2007 University of Cambridge. Permissions and limitations regarding this software are subject to terms available at
http:// www.pcre.org/license.txt.

This product includes software copyright © 2007 The Eclipse Foundation. All Rights Reserved. Permissions and limitations regarding this software are subject to terms
available at http:// www.eclipse.org/org/documents/epl-v10.php and at http://www.eclipse.org/org/documents/edl-v10.php.
This product includes software licensed under the terms at http://www.tcl.tk/software/tcltk/license.html, http://www.bosrup.com/web/overlib/?License, http://
www.stlport.org/doc/ license.html, http://asm.ow2.org/license.html, http://www.cryptix.org/LICENSE.TXT, http://hsqldb.org/web/hsqlLicense.html, http://
httpunit.sourceforge.net/doc/ license.html, http://jung.sourceforge.net/license.txt , http://www.gzip.org/zlib/zlib_license.html, http://www.openldap.org/software/release/
license.html, http://www.libssh2.org, http://slf4j.org/license.html, http://www.sente.ch/software/OpenSourceLicense.html, http://fusesource.com/downloads/license-
agreements/fuse-message-broker-v-5-3- license-agreement; http://antlr.org/license.html; http://aopalliance.sourceforge.net/; http://www.bouncycastle.org/licence.html;
http://www.jgraph.com/jgraphdownload.html; http://www.jcraft.com/jsch/LICENSE.txt; http://jotm.objectweb.org/bsd_license.html; . http://www.w3.org/Consortium/Legal/
2002/copyright-software-20021231; http://www.slf4j.org/license.html; http://nanoxml.sourceforge.net/orig/copyright.html; http://www.json.org/license.html; http://
forge.ow2.org/projects/javaservice/, http://www.postgresql.org/about/licence.html, http://www.sqlite.org/copyright.html, http://www.tcl.tk/software/tcltk/license.html, http://
www.jaxen.org/faq.html, http://www.jdom.org/docs/faq.html, http://www.slf4j.org/license.html; http://www.iodbc.org/dataspace/iodbc/wiki/iODBC/License; http://
www.keplerproject.org/md5/license.html; http://www.toedter.com/en/jcalendar/license.html; http://www.edankert.com/bounce/index.html; http://www.net-snmp.org/about/
license.html; http://www.openmdx.org/#FAQ; http://www.php.net/license/3_01.txt; http://srp.stanford.edu/license.txt; http://www.schneier.com/blowfish.html; http://
www.jmock.org/license.html; http://xsom.java.net; http://benalman.com/about/license/; https://github.com/CreateJS/EaselJS/blob/master/src/easeljs/display/Bitmap.js;
http://www.h2database.com/html/license.html#summary; http://jsoncpp.sourceforge.net/LICENSE; http://jdbc.postgresql.org/license.html; http://
protobuf.googlecode.com/svn/trunk/src/google/protobuf/descriptor.proto; https://github.com/rantav/hector/blob/master/LICENSE; http://web.mit.edu/Kerberos/krb5-
current/doc/mitK5license.html; http://jibx.sourceforge.net/jibx-license.html; https://github.com/lyokato/libgeohash/blob/master/LICENSE; https://github.com/hjiang/jsonxx/
blob/master/LICENSE; https://code.google.com/p/lz4/; https://github.com/jedisct1/libsodium/blob/master/LICENSE; http://one-jar.sourceforge.net/index.php?
page=documents&file=license; https://github.com/EsotericSoftware/kryo/blob/master/license.txt; http://www.scala-lang.org/license.html; https://github.com/tinkerpop/
blueprints/blob/master/LICENSE.txt; http://gee.cs.oswego.edu/dl/classes/EDU/oswego/cs/dl/util/concurrent/intro.html; https://aws.amazon.com/asl/; https://github.com/
twbs/bootstrap/blob/master/LICENSE; and https://sourceforge.net/p/xmlunit/code/HEAD/tree/trunk/LICENSE.txt.
This product includes software licensed under the Academic Free License (http://www.opensource.org/licenses/afl-3.0.php), the Common Development and Distribution
License (http://www.opensource.org/licenses/cddl1.php) the Common Public License (http://www.opensource.org/licenses/cpl1.0.php), the Sun Binary Code License
Agreement Supplemental License Terms, the BSD License (http:// www.opensource.org/licenses/bsd-license.php), the new BSD License (http://opensource.org/
licenses/BSD-3-Clause), the MIT License (http://www.opensource.org/licenses/mit-license.php), the Artistic License (http://www.opensource.org/licenses/artistic-
license-1.0) and the Initial Developer’s Public License Version 1.0 (http://www.firebirdsql.org/en/initial-developer-s-public-license-version-1-0/).

This product includes software copyright © 2003-2006 Joe WaInes, 2006-2007 XStream Committers. All rights reserved. Permissions and limitations regarding this
software are subject to terms available at http://xstream.codehaus.org/license.html. This product includes software developed by the Indiana University Extreme! Lab.
For further information please visit http://www.extreme.indiana.edu/.

This product includes software Copyright (c) 2013 Frank Balluffi and Markus Moeller. All rights reserved. Permissions and limitations regarding this software are subject
to terms of the MIT license.

See patents at https://www.informatica.com/legal/patents.html.

DISCLAIMER: Informatica LLC provides this documentation "as is" without warranty of any kind, either express or implied, including, but not limited to, the implied
warranties of noninfringement, merchantability, or use for a particular purpose. Informatica LLC does not warrant that this software or documentation is error free. The
information provided in this software or documentation may include technical inaccuracies or typographical errors. The information in this software and documentation is
subject to change at any time without notice.

NOTICES

This Informatica product (the "Software") includes certain drivers (the "DataDirect Drivers") from DataDirect Technologies, an operating company of Progress Software
Corporation ("DataDirect") which are subject to the following terms and conditions:

1. THE DATADIRECT DRIVERS ARE PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
2. IN NO EVENT WILL DATADIRECT OR ITS THIRD PARTY SUPPLIERS BE LIABLE TO THE END-USER CUSTOMER FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, CONSEQUENTIAL OR OTHER DAMAGES ARISING OUT OF THE USE OF THE ODBC DRIVERS, WHETHER OR NOT
INFORMED OF THE POSSIBILITIES OF DAMAGES IN ADVANCE. THESE LIMITATIONS APPLY TO ALL CAUSES OF ACTION, INCLUDING, WITHOUT
LIMITATION, BREACH OF CONTRACT, BREACH OF WARRANTY, NEGLIGENCE, STRICT LIABILITY, MISREPRESENTATION AND OTHER TORTS.

Part Number: PWX-IPC100-0001

Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Informatica Resources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Informatica My Support Portal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Informatica Documentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica Product Availability Matrixes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica Web Site. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica How-To Library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica Knowledge Base. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica Support YouTube Channel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Informatica Marketplace. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Informatica Velocity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Informatica Global Customer Support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Part I: Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Chapter 1: PowerExchange Interfaces for PowerCenter. . . . . . . . . . . . . . . . . . . . 14

PowerExchange Interfaces for PowerCenter Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
PowerExchange Client for PowerCenter (PWXPC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Bulk Extraction Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
CDC Batch and Continuous Extraction Modes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
CDC Real-time Extraction Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
PowerExchange ODBC Drivers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Part II: PowerExchange Client for PowerCenter (PWXPC). . . . . . . . . . . . . . . . . . . . . . 22

Chapter 2: Installing PowerExchange Client for PowerCenter (PWXPC). . . . . . . 23

PWXPC Installation Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Installing and Configuring PWXPC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Configuration Step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Upgrade Steps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Modifying the PowerExchange Configuration Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Adding PowerExchange Nodes on the PowerCenter Client Machine. . . . . . . . . . . . . . . . . . 25
Adding PowerExchange Nodes on the PowerCenter Integration Service Machine. . . . . . . . . 25
Adding PowerExchange Nodes in the dbmover.cfg File. . . . . . . . . . . . . . . . . . . . . . . . . . 25
PowerExchange Interoperability with PowerCenter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Chapter 3: Working with Mappings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Working with Mappings Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Source and Target Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Working with Relational Source and Target Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4 Table of Contents
 Working with DB2 Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Working with Microsoft SQL Server Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Working with Oracle Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Editing Relational Source and Target Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Working with Nonrelational Source and Target Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Importing Nonrelational Source Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Importing Nonrelational Target Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Viewing Nonrelational Source and Target Definition Details. . . . . . . . . . . . . . . . . . . . . . . . 41
Editing Nonrelational Source and Target Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Working with Extraction Map Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Viewing Extraction Map Definition Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Editing an Extraction Map Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Previewing PowerExchange Data in Designer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Previewing PowerExchange Relational Source or Target Data. . . . . . . . . . . . . . . . . . . . . . 48
Previewing PowerExchange Nonrelational Source or Target Data. . . . . . . . . . . . . . . . . . . . 49
Previewing PowerExchange Change Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Group Source Processing in PowerExchange. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Group Source Processing in Bulk Data Movement Sessions with Nonrelational Data Sources. . 51
Group Source Processing in CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Multiple-Record Writes to Nonrelational Targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Group Processing with Sequencing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Sequencing Support in Source and Target Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Working with Source Qualifiers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Transformation Datatypes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Creating Source Qualifiers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Editing Source Qualifier Transformations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Lookup Transformations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Usage Considerations for PowerExchange Sources and Targets. . . . . . . . . . . . . . . . . . . . 57
Limitations on Nonrelational Uncached Lookups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Configuring Lookups for IMS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Configuring Lookups for CDC Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Stored Procedure Transformations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Usage Considerations for Stored Procedure Transformations. . . . . . . . . . . . . . . . . . . . . . 59
Implementing a Stored Procedure Transformation in a Mapping. . . . . . . . . . . . . . . . . . . . . 60

Chapter 4: Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Connections Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Connection Types for Extracting Source Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Connection Types for Loading Target Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Configuring PowerExchange Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
DB2 Relational Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
DB2 CDC Application Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
NRDB Batch Application Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Table of Contents 5
 NRDB CDC Application Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
NRDB Lookup Relational Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Microsoft SQL Server Relational Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Microsoft SQL Server CDC Application Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Oracle Relational Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Oracle CDC Application Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Configuring Connection Attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Common Connection Attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Batch Application and Relational Connection Attributes. . . . . . . . . . . . . . . . . . . . . . . . . 115
CDC Application Connection Attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Connection Attributes for CDC Commit Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Examples of Using CDC Commit Processing Attributes. . . . . . . . . . . . . . . . . . . . . . . . . 127
Connection Attributes for Multiple-Record Writes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Asynchronous Network Communication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Asynchronous Read and Write Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Timeout Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Configuring Connection and Network Operation Timeouts. . . . . . . . . . . . . . . . . . . . . . . . 131

Chapter 5: Working with Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

Working with Sessions Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Extracting Data in Batch Mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Data Extraction from Multiple Sequential or VSAM ESDS Files. . . . . . . . . . . . . . . . . . . . . 134
Overrides for the IMS Access Method and Related Properties. . . . . . . . . . . . . . . . . . . . . 134
Considerations for Sessions that Perform Multiple-Record Writes. . . . . . . . . . . . . . . . . . . 135
Configuring Batch Mode Sessions for Nonrelational Sources. . . . . . . . . . . . . . . . . . . . . . 138
Configuring Session Properties for Relational Batch Mode Sessions. . . . . . . . . . . . . . . . . 144
Extracting CDC Data in Change and Real-time Modes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
PowerExchange Extraction Maps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Data Extraction from Multiple CDC Sources in a Session. . . . . . . . . . . . . . . . . . . . . . . . 147
Session Properties for Change and Real-time Mode Sessions. . . . . . . . . . . . . . . . . . . . . 147
Asterisk Wildcard in i5/OS Library or File Overrides. . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Treating Source Rows as Updates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Loading Data to PowerExchange Targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Configuring Sessions to Load Data to Relational Targets. . . . . . . . . . . . . . . . . . . . . . . . 153
Configuring Sessions to Load Data to Nonrelational Targets. . . . . . . . . . . . . . . . . . . . . . 153
Constraint-Based Loading for Relational Targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Configuring Constraint-Based Loading for Relational Targets. . . . . . . . . . . . . . . . . . . . . . 157
Variables for Session and Connection Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Workflows that Connect to a PowerExchange Listener Service. . . . . . . . . . . . . . . . . . . . . . . . 159
Configuring PowerExchange Services for Workflows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Pipeline Partitioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Commit Processing for Bulk Data Movement Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Commit Processing for CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

6 Table of Contents
Considerations for Pre- and Post-Session SQL Commands. . . . . . . . . . . . . . . . . . . . . . . . . . 161
Preserving Low Values in Source Character Fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

Chapter 6: Restart and Recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

Restart and Recovery Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Understanding PWXPC Restart and Recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Session Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Recovery Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Recovery State Table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Recovery State File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Restart Token File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Determining the Restart Point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Initializing and Running CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Ending CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Creating Recovery Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Recovery Table Creation with PowerExchange Targets. . . . . . . . . . . . . . . . . . . . . . . . . 172
Creating the Recovery Tables Manually. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Configuring the Restart Token File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Restart Token File Control Statement Syntax. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Restart Token File Control Statement Parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
Restart Token File Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
PWXPC Restart and Recovery Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Enabling Session Recovery for a CDC Session. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Configuring CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Application Names. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Generating Current Restart Tokens for CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . 179
Starting CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
Stopping CDC Sessions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Changing CDC Sessions and Restarting from the Point of Interruption. . . . . . . . . . . . . . . . 183
Recovering from CDC Session Failures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Managing Session Log and Restart Token File History. . . . . . . . . . . . . . . . . . . . . . . . . . 185

Chapter 7: Flexible Target Key Transformation. . . . . . . . . . . . . . . . . . . . . . . . . . 187

Flexible Target Key Transformation Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Configuring PowerExchange Extraction Maps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Configuring Mappings with Flexible Target Key Transformations. . . . . . . . . . . . . . . . . . . . . . . 189
Configuring Flexible Target Key Transformations with Multiple Sources. . . . . . . . . . . . . . . . . . 190

Part III: PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Chapter 8: Installing PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Installing PowerExchange ODBC Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Modifying the PowerExchange Configuration Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Table of Contents 7
 Creating ODBC Data Sources. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Creating an ODBC Data Source on Windows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Creating a PowerExchange ODBC Data Source on Linux or UNIX. . . . . . . . . . . . . . . . . . 195

Chapter 9: Working with Mappings for PowerExchange ODBC. . . . . . . . . . . . . 196

Working with Mappings for PowerExchange ODBC Overview. . . . . . . . . . . . . . . . . . . . . . . . 196
Working with Source and Target Definitions for PowerExchange Batch (ODBC). . . . . . . . . . . . . 197
Importing a Source or Target Definition for PowerExchange ODBC. . . . . . . . . . . . . . . . . . 197
Working with Source Definitions for PowerExchange Change or Real-time (ODBC). . . . . . . . . . 198
Importing a Change Data Capture Source Definition for PowerExchange ODBC. . . . . . . . . 198

Chapter 10: Configuring Connections for PowerExchange ODBC. . . . . . . . . . . 200

Configuring Connections for PowerExchange ODBC Overview. . . . . . . . . . . . . . . . . . . . . . . . 200
Working with Connections for PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Connection Types for Extracting Source Data for PowerExchange ODBC. . . . . . . . . . . . . . 200
Connection Types for Loading Target Data for PowerExchange ODBC. . . . . . . . . . . . . . . 201
Configuring Attributes for Connections for PowerExchange ODBC. . . . . . . . . . . . . . . . . . 202

Chapter 11: Working with Sessions for PowerExchange ODBC. . . . . . . . . . . . . 203

Working with Sessions for PowerExchange ODBC Overview. . . . . . . . . . . . . . . . . . . . . . . . . 203
Pipeline Partitioning for PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Extracting Data from PowerExchange in Batch Mode (ODBC). . . . . . . . . . . . . . . . . . . . . . . . 204
Configuring Properties for Nonrelational Batch Mode Sessions (ODBC). . . . . . . . . . . . . . . 204
Configuring Properties for Relational Batch Mode Sessions (ODBC). . . . . . . . . . . . . . . . . 205
Extracting Data from PowerExchange in Change and Real-time Mode (ODBC). . . . . . . . . . . . . 205
Configuring Properties for Change and Real-time Mode Sessions (ODBC). . . . . . . . . . . . . 206
Configuring a Session to Load Data to PowerExchange Targets (ODBC). . . . . . . . . . . . . . . . . 206

Chapter 12: PowerExchange Restart and Recovery. . . . . . . . . . . . . . . . . . . . . . 207

PowerExchange Restart and Recovery Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Restart and Recovery with PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Managing Restart Tokens with PowerExchange ODBC. . . . . . . . . . . . . . . . . . . . . . . . . . 208

Appendix A: PowerExchange Interfaces for PowerCenter Tips. . . . . . . . . . . . . . . . 209

Organizing Sources by Map Type in the Designer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Filtering Source Data with PWXPC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Filter Override Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
DTLREXE to Submit MVS Batch Jobs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Empty Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Creating Empty Files in a PowerCenter Session. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Empty File - Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

8 Table of Contents
Appendix B: Datatypes and Code Pages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Datatypes and Code Pages Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
PowerExchange Nonrelational Datatypes and Transformation Datatypes. . . . . . . . . . . . . . . . . 217
Restrictions on Relational Datatypes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
LOB Datatypes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
DB2 for z/OS TIMESTAMP Datatype. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Reading and Writing Binary Data in PowerExchange Client for PowerCenter . . . . . . . . . . . . . . 221
Code Pages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

Appendix C: PowerExchange Interfaces for PowerCenter Troubleshooting. . . . 222

Troubleshooting for PowerExchange Interfaces for PowerCenter. . . . . . . . . . . . . . . . . . . . . . 222

Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

Table of Contents 9
Preface
The PowerExchange Interfaces for PowerCenter publication describes the Informatica PowerExchange Client
for PowerCenter (PWXPC) and ODBC interfaces between PowerExchange and PowerCenter. This
publication is intended for developers and administrators who are responsible for creating, running, and
administering workflows and sessions that interface with PowerExchange.

This publication assumes that you have knowledge of relational database concepts and the database engines
and nonrelational files in your environment. This guide also assumes that you are familiar with the basic
operation of PowerExchange and PowerCenter.

This publication supplements the information in the PowerCenter library, including the PowerCenter Designer
Guide, PowerCenter Transformation Guide, PowerCenter Workflow Basics Guide, PowerCenter Advanced
Workflow Guide, and Informatica Administrator Guide.

This publication applies to the following PowerExchange products, including PowerExchange change data
capture (CDC) where applicable:

• PowerExchange for Adabas®

• PowerExchange for CA Datacom®
• PowerExchange for CA IDMS™
• PowerExchange for DB2® for i5/OS®
• PowerExchange for DB2 for Linux®, UNIX®, and Windows®
• PowerExchange for DB2 for z/OS®
• PowerExchange for IMS™
• PowerExchange for Flat Files on Linux, UNIX, and Windows
• PowerExchange for Oracle
• PowerExchange for SQL Server®
• PowerExchange for VSAM

Informatica Resources

Informatica My Support Portal

As an Informatica customer, the first step in reaching out to Informatica is through the Informatica My Support
Portal at https://mysupport.informatica.com. The My Support Portal is the largest online data integration
collaboration platform with over 100,000 Informatica customers and partners worldwide.

10
 As a member, you can:

• Access all of your Informatica resources in one place.

• Review your support cases.
• Search the Knowledge Base, find product documentation, access how-to documents, and watch support
videos.
• Find your local Informatica User Group Network and collaborate with your peers.

Informatica Documentation
The Informatica Documentation team makes every effort to create accurate, usable documentation. If you
have questions, comments, or ideas about this documentation, contact the Informatica Documentation team
through email at [email protected]. We will use your feedback to improve our
documentation. Let us know if we can contact you regarding your comments.

The Documentation team updates documentation as needed. To get the latest documentation for your
product, navigate to Product Documentation from https://mysupport.informatica.com.

Informatica Product Availability Matrixes

Product Availability Matrixes (PAMs) indicate the versions of operating systems, databases, and other types
of data sources and targets that a product release supports. You can access the PAMs on the Informatica My
Support Portal at https://mysupport.informatica.com.

Informatica Web Site

You can access the Informatica corporate web site at https://www.informatica.com. The site contains
information about Informatica, its background, upcoming events, and sales offices. You will also find product
and partner information. The services area of the site includes important information about technical support,
training and education, and implementation services.

Informatica How-To Library

As an Informatica customer, you can access the Informatica How-To Library at
https://mysupport.informatica.com. The How-To Library is a collection of resources to help you learn more
about Informatica products and features. It includes articles and interactive demonstrations that provide
solutions to common problems, compare features and behaviors, and guide you through performing specific
real-world tasks.

Informatica Knowledge Base

As an Informatica customer, you can access the Informatica Knowledge Base at
https://mysupport.informatica.com. Use the Knowledge Base to search for documented solutions to known
technical issues about Informatica products. You can also find answers to frequently asked questions,
technical white papers, and technical tips. If you have questions, comments, or ideas about the Knowledge
Base, contact the Informatica Knowledge Base team through email at [email protected].

Informatica Support YouTube Channel

You can access the Informatica Support YouTube channel at http://www.youtube.com/user/INFASupport. The
Informatica Support YouTube channel includes videos about solutions that guide you through performing
specific tasks. If you have questions, comments, or ideas about the Informatica Support YouTube channel,

Preface 11
 contact the Support YouTube team through email at [email protected] or send a tweet to
@INFASupport.

Informatica Marketplace
The Informatica Marketplace is a forum where developers and partners can share solutions that augment,
extend, or enhance data integration implementations. By leveraging any of the hundreds of solutions
available on the Marketplace, you can improve your productivity and speed up time to implementation on
your projects. You can access Informatica Marketplace at http://www.informaticamarketplace.com.

Informatica Velocity
You can access Informatica Velocity at https://mysupport.informatica.com. Developed from the real-world
experience of hundreds of data management projects, Informatica Velocity represents the collective
knowledge of our consultants who have worked with organizations from around the world to plan, develop,
deploy, and maintain successful data management solutions. If you have questions, comments, or ideas
about Informatica Velocity, contact Informatica Professional Services at [email protected].

Informatica Global Customer Support

You can contact a Customer Support Center by telephone or through the Online Support.

Online Support requires a user name and password. You can request a user name and password at
http://mysupport.informatica.com.

The telephone numbers for Informatica Global Customer Support are available from the Informatica web site
at http://www.informatica.com/us/services-and-training/support-services/global-support-centers/.

12 Preface
Part I: Introduction
This part contains the following chapter:

• PowerExchange Interfaces for PowerCenter, 14

13
CHAPTER 1

PowerExchange Interfaces for

PowerCenter
This chapter includes the following topics:

• PowerExchange Interfaces for PowerCenter Overview, 14

• PowerExchange Client for PowerCenter (PWXPC), 15
• PowerExchange ODBC Drivers, 20

PowerExchange Interfaces for PowerCenter

Overview
You can use the following interfaces to extract and load data through PowerExchange when you use
PowerCenter:

• PowerExchange Client for PowerCenter (PWXPC). Part I describes PWXPC. You can use it to extract
and load data through PowerExchange for a variety of datatypes on a variety of platforms. PWXPC is fully
integrated into PowerCenter.
• PowerExchange ODBC. Part II describes the PowerExchange ODBC interface. You can use
PowerExchange ODBC connections with PowerCenter to extract and load data through PowerExchange
for a variety of datatypes on a variety of platforms.

Note: When connecting to PowerExchange, Informatica recommends that you use PWXPC instead of
PowerExchange ODBC. PWXPC has additional functionality as well as improved performance and superior
CDC recovery and restart.

14
 The following table compares the interface functionality of the PowerExchange Client for PowerCenter and
PowerExchange ODBC:

PWXPC ODBC Function Description

Yes Yes Extracts bulk data and changed PowerExchange extracts bulk data for relational and
data nonrelational sources and changed data for CDC
sources.

Yes No Extracts data for multiple PowerExchange uses group source to extract
sources or records in a single changed data for multiple data sources or bulk data
pass for multiple record types in a VSAM and sequential
files in a single pass.

Yes No Saves target data and CDC CDC restart information is stored in the same
restart information in a single database as the relational target table or MQ queue.
commit The restart information is updated in the same
commit as the target data providing guaranteed
restart and recovery for CDC data.

Yes No Uses PowerCenter graceful stop PowerCenter stops real-time sessions after all data
for real-time sessions in the pipeline is written to the targets.

Yes No Uses the change indicator to Each change record indicates whether it is an insert,
determine the type of change update, or delete. When the change indicator is
record used, an Update Strategy transformation is not
required to process inserts, updates, and deletes.

Yes No Creates source definitions from Extraction maps contain the PowerExchange auto-
PowerExchange extraction generated columns minimizing modification of the
maps source definition in Designer.

Yes No Uses the file name in the PowerCenter source definition can specify the file
PowerCenter source definition name and override the file name specified in the
PowerExchange data map.

PowerExchange Client for PowerCenter (PWXPC)

PowerExchange Client for PowerCenter is installed with PowerCenter and integrates PowerExchange and
PowerCenter to extract relational, nonrelational, and changed data. PWXPC also loads relational and
nonrelational data in batch mode.

Use the following modes to extract relational and nonrelational data:

• Bulk data movement

• Change data capture (CDC) real time
• CDC batch extraction mode from condense files
• CDC continuous extraction mode from condense files

PowerExchange Client for PowerCenter (PWXPC) 15

 The following table lists the database types that PowerExchange Client for PowerCenter can access to
extract data or to load data:

Database Type Bulk Mode CDC Real-time CDC Batch CDC Continuous
Extract/Load Extraction Mode Extraction Mode Extraction Mode

Adabas Yes/Yes Yes Yes No

Datacom Yes/No Yes Yes No

DB2 for z/OS Yes/Yes Yes Yes No

DB2 for i5/OS Yes/Yes Yes Yes No

DB2 for Linux, UNIX, Yes/Yes Yes Yes Yes

and Windows

IDMS Yes/No Yes Yes No

IMS Yes/Yes Yes Yes No

MSSQL Yes/Yes Yes Yes Yes

Oracle Yes/Yes Yes Yes Yes

Sequential files Yes/Yes No No No

VSAM Yes/Yes Yes Yes No

Note: For VSAM ESDS
and RRDS data sets,
only inserts are
allowed. For VSAM
KSDS data sets,
inserts, updates, and
deletes are allowed.

PowerExchange group source reads data for data sets and files with multiple record types or multiple CDC
data sources in a single pass. PWXPC uses PowerExchange group source to extract changed data from the
change stream. PWXPC also uses group source to extract data for multiple record types from VSAM data
sets, sequential files, and IMS unload data sets. As a result, PWXPC connections can process data faster
than PowerExchange ODBC connections and reduce PowerExchange resource consumption on the source
or extraction platform.

The following table lists the PowerExchange database types that read sources in a single pass during
extraction:

Database Type Bulk Extraction CDC Real-time CDC Batch CDC Continuous
Mode Extraction Mode Extraction Mode Extraction Mode

Adabas No Yes Yes No

Datacom No Yes Yes No

DB2 for z/OS No Yes Yes No

DB2 for i5/OS No Yes Yes No

16 Chapter 1: PowerExchange Interfaces for PowerCenter

 Database Type Bulk Extraction CDC Real-time CDC Batch CDC Continuous
Mode Extraction Mode Extraction Mode Extraction Mode

DB2 for Linux, UNIX, No Yes Yes Yes

and Windows

IDMS No Yes Yes No

IMS Yes Yes Yes No

Note: IMS unload
data sets only.

Microsoft SQL Server No Yes Yes Yes

Oracle No Yes Yes Yes

Sequential files Yes No No No

VSAM Yes Yes Yes No

Related Topics:
• “Working with Source Qualifiers” on page 56

Bulk Extraction Mode

Use PWX batch application and relational connections to extract and load data for relational databases and
nonrelational data sets and files through PowerExchange. PWXPC connects to PowerExchange through the
PowerExchange Call Level Interface (SCLI).

When you use PWXPC, you can extract all records for multiple record types in a single pass for IMS unload
data sets, VSAM data sets, and sequential files. In contrast, PowerExchange ODBC connections read a
single record type at a time, which requires multiple passes of the data.

PowerExchange Client for PowerCenter (PWXPC) 17

 The following figure shows the data flow of source data from PowerExchange through PWXPC and
PowerCenter to the target tables:

CDC Batch and Continuous Extraction Modes

PowerExchange Condense captures changes for relational tables and nonrelational data sets and files from
the change stream and writes the changed data to condense files. The change stream contains the source
data changes in time sequence order. As a result, the changes in a single unit of work are not contiguous in
the change stream. The UOW Cleanser reconstructs the units of work into complete and consecutive units of
work based on transaction end time. PowerExchange Condense writes the reconstructed units of work to
condense files after which you can use extract the changed data.

PWXPC extracts changed data from PowerExchange condense files in two different modes:

• Batch extraction mode from condense files. Use PWX CDC Change connections to extract changed
data from condense files in batch extraction mode. PWXPC uses the PowerExchange CAPX access
method when processing data with CDC Change connections. PowerExchange stops the extraction after
the data from all condense files is read.
• Continuous extraction mode from condense files. Use PWX CDC Real Time connections for Linux,
UNIX, and Windows sources to extract changed data from condense files in continuous extraction mode.
PWXPC uses the PowerExchange CAPXRT access method when processing data with CDC Real Time
connections. PowerExchange runs the extraction until stopped. For more information, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.
PWXPC connects to PowerExchange through the PowerExchange Call Level Interface (SCLI).
PowerExchange reads the changed data from each condense file once for all sources in the mapping in a
single pass.

18 Chapter 1: PowerExchange Interfaces for PowerCenter

 The following figure shows the data flow of condensed changed data from PowerExchange through PWXPC
and PowerCenter to the target tables:

CDC Real-time Extraction Mode

You can extract changed data for relational tables and nonrelational data sets and files directly from the
change stream in real time. The change stream contains the source data changes in time sequence order. As
a result, the changes in a single unit of work are not contiguous in the change stream. The UOW Cleanser
reconstructs the units of work into complete and consecutive units of work based on transaction end time.
The UOW Cleanser is incorporated into UDB and SQL Server CDC but is a separate task for all other
PowerExchange CDC sources.

Use PWX CDC Real Time application connections to extract changed data for relational tables or
nonrelational data sets and files directly from the change stream. PWXPC connects to PowerExchange
through the PowerExchange Call Level Interface (SCLI) and specifying the PowerExchange CAPXRT access
method. PowerExchange reads the changed data from the change stream for all sources in the mapping in a
single pass. PWXPC real-time sessions run for a specified period or continuously until stopped.

PowerExchange Client for PowerCenter (PWXPC) 19

 The following figure shows the data flow of changed data from PowerExchange through PWXPC and
PowerCenter to the target tables:

PowerExchange ODBC Drivers

PowerExchange provides a thin ODBC driver that you can use with PowerCenter. The PowerExchange
Integration Service uses PowerExchange ODBC to connect to PowerExchange either locally or remotely.
With PowerExchange ODBC, you can extract and load relational and nonrelational data. You can also extract
changed data.

Use the following modes to extract relational and nonrelational data:

• Batch. PowerExchange ODBC extracts and loads data from a relational table or nonrelational file through
PowerExchange. You can read multiple record VSAM data sets and sequential files through the ODBC
interface with multiple passes of the data to read all record types.
• Change Data Capture (CDC) batch extraction mode from condense files. PowerExchange ODBC
extracts changed data from condense files through PowerExchange reading all of the changes captured in
condense files since the last extraction session. PowerExchange ODBC reads the changed data once for
each source in the mapping resulting in multiple passes of the condense files. The extraction session
ends when all captured changes are read. PowerExchange maintains restart information in the CDEP file
on the source machine. PowerExchange ODBC has limited restart capability.
• CDC real time. PowerExchange ODBC extracts changed data in real time from the change stream by
using one pass of the data for each source in the mapping. You can run real-time extractions for a
specified time period or continuously until stopped. PowerExchange maintains restart information in the
CDEP file on the source machine. PowerExchange ODBC has limited restart capability.

20 Chapter 1: PowerExchange Interfaces for PowerCenter

The following table lists the PowerExchange ODBC extract and load capabilities:

Database Type Batch Mode Batch Mode Load CDC Batch CDC Real-time
Extract Extraction Mode Extraction Mode

Adabas Yes Yes Yes Yes

Datacom Yes No Yes Yes

DB2 for z/OS Yes Yes Yes Yes

DB2 for i5/OS Yes Yes Yes Yes

DB2 for Linux, Yes Yes Yes Yes

UNIX, and
Windows

IDMS Yes No Yes Yes

IMS Yes Yes Yes Yes

MSSQL No Yes Yes Yes

Oracle No Yes Yes Yes

Sequential/flat files Yes Yes No No

VSAM Yes Yes Yes Yes

Note: For VSAM
ESDS and RRDS
data sets, only
inserts are allowed.
For VSAM KSDS
data sets, inserts,
updates, and
deletes are
allowed.

PowerExchange ODBC Drivers 21

Part II: PowerExchange Client for
PowerCenter (PWXPC)
This part contains the following chapters:

• Installing PowerExchange Client for PowerCenter (PWXPC), 23

• Working with Mappings, 27
• Connections, 61
• Working with Sessions, 133
• Restart and Recovery, 163
• Flexible Target Key Transformation, 187

22
CHAPTER 2

Installing PowerExchange Client

for PowerCenter (PWXPC)
This chapter includes the following topics:

• PWXPC Installation Overview, 23

• Installing and Configuring PWXPC, 23
• Modifying the PowerExchange Configuration Files, 25
• PowerExchange Interoperability with PowerCenter, 26

PWXPC Installation Overview

The PowerExchange Client for PowerCenter (PWXPC) is a native PowerCenter plug-in that is installed with
PowerCenter.

To use PWXPC, you must also install PowerExchange on the PowerCenter Client and PowerCenter
Integration Service machines.

Note: If the appropriate version of PowerExchange is not installed and available on the PowerCenter Client
platform, the Import from PowerExchange dialog box will not function.

Installing and Configuring PWXPC

To install PWXPC, complete the PowerCenter installation steps for the PowerCenter Integration Service and
PowerCenter Client.

Also, install and configure PowerExchange on the machines where the PowerCenter Integration Service and
PowerCenter Client run. To use PWXPC, you must complete some PowerExchange configuration tasks.

Configuration Step
To configure PWXPC on the PowerCenter Integration Service and PowerCenter Client, you must configure
PowerExchange dbmover.cfg configuration files on those PowerCenter nodes.

In these dbmover.cfg files, add NODE statements that point to the PowerExchange Listeners to which you
want PWXPC to connect.

23
 For more information, see the PowerExchange Reference Manual.

Related Topics:
• “Modifying the PowerExchange Configuration Files” on page 25

Upgrade Steps
When you upgrade PowerCenter, you might need to complete some additional upgrade tasks for PWXPC.

For PWXPC, complete the following tasks:

• Re-register the PWXPC plug-in and Flexible Target Key transformation plug-in.
• If you use PowerExchange Change Data Capture (CDC), configure the restart token file to establish an
extraction restart point for CDC sessions.

Registering Plug-ins the PowerCenter Repository

When you upgrade PowerCenter from releases earlier than 8.5, you must perform a repository upgrade. The
upgrade process registers native plug-ins.

If you do not perform a repository upgrade as a part of a PowerCenter upgrade, you must register the
PWXPC and Flexible Target Key transformation plug-ins.

To register the PWXPC and Flexible Target Key transformation plug-ins:

1. In the Administrator tool, set the repository operating mode to Exclusive.

2. Locate the pmpwxclnt.xml and pmflxtgtkeytransform.xml files in the native directory at one of the
following paths:
On Windows:
%INFA_HOME%/server/bin/native
On Linux or UNIX:
$INFA_HOME/server/bin/native
The %INFA_HOME% and $INFA_HOME variables represent the PowerCenter installation directory.
3. To connect to the repository, use the pmrep command line program to issue the following command:
connect -r repository_name -d domain_name -n repository_user
–x repository_password
4. To register the plug-ins, use pmrep to issue the following commands:
• For a repository on a Windows machine:
registerplugin -i %INFA_HOME%\server\bin\native\pmpwxclnt.xml -e –N
registerplugin -i %INFA_HOME%\server\bin\native\pmflxtgtkeytransform.xml -e –N
• For a repository on a Linux or UNIX machine:
registerplugin -i $INFA_HOME/server/bin/native/pmpwxclnt.xml -e –N
registerplugin -i $INFA_HOME/server/bin/native/pmflxtgtkeytransform.xml -e –N
5. Exit the pmrep command.
6. Use the Administrator tool to set the repository operating mode to Normal.

Establishing Restart Points for CDC Sessions

If you upgrade from PowerCenter 8.6, you can warm start CDC sessions after the PowerCenter upgrade.

If you upgrade from a PowerCenter release earlier than 8.6 and use enhanced restart for CDC sessions, you
must cold start all CDC sessions.

24 Chapter 2: Installing PowerExchange Client for PowerCenter (PWXPC)

Modifying the PowerExchange Configuration Files
You must define nodes for the PowerExchange Listeners in the PowerExchange configuration file,
dbmover.cfg, on the PowerCenter Integration Service and Client machines. The NODE statements specify
the information that PowerExchange uses to communicate with Listeners that access the source or target
databases or files.

Note: You can also use PowerExchange local mode if the data resides on the same machine as the
PowerCenter Integration Service. In local mode, a PowerExchange Listener is not required. If you use local
mode, specify local in the Location property in PWXPC connections. You do not need to update the
PowerExchange dbmover.cfg file.

Local mode is not available on 32-bit Windows systems.

Adding PowerExchange Nodes on the PowerCenter Client

Machine
PWXPC connects to a PowerExchange Listener to import source and target definitions. PWXPC retrieves the
metadata from the machine indicated in the NODE statement or retrieves the metadata locally in local mode.

Update the dbmover.cfg file on the PowerCenter Client machine to provide the appropriate NODE statements
for the PowerExchange Listeners to which you need to connect.

The nodes you add to the dbmover.cfg file are specified in the Location field in Import from PowerExchange
dialog box.

Adding PowerExchange Nodes on the PowerCenter Integration

Service Machine
PWXPC connects to the appropriate PowerExchange Listener to extract data from a source or load data to a
target.

Update the dbmover.cfg file on the PowerCenter Integration Service machine to provide the appropriate
NODE statements for the PowerExchange Listeners to which you need to connect.

The nodes you add to the dbmover.cfg file on the PowerCenter Integration Service machine are specified in
the Location value of the PWXPC connections.

Adding PowerExchange Nodes in the dbmover.cfg File

Use the following procedure to add PowerExchange nodes in the dbmover.cfg file:

1. Locate the dbmover.cfg file in the PowerExchange root directory.

2. Open the file with a text editor.
3. Create a node for each PowerExchange Listener that you want to register. Use the following format:
NODE=(node_name,TCPIP,hostname_or_ipaddress,port_number)
The node name is a logical name used to reference the PowerExchange Listener. The
hostname_or_ipaddress and port_number are the host name or IP address and the port number of the
PowerExchange Listener.
4. Save your changes.
The following example shows several nodes in a dbmover.cfg file:
/********************************************************************/
/* PowerExchange Configuration File

Modifying the PowerExchange Configuration Files 25

 /********************************************************************/
NODE=(AS400_DB2,TCPIP,AS400_1,2480)
NODE=(MVS1_VSAM,TCPIP,MVS1,5539)
NODE=(MVS2_DB2,TCPIP,10.3.4.5,5538)

PowerExchange Interoperability with PowerCenter

PowerCenter includes the PowerExchange Client for PowerCenter (PWXPC). To use PWXPC to connect to
PowerExchange, you must verify that the PowerCenter and PowerExchange versions are compatible.

The following table describes the supported combinations of PowerExchange (PWX) and PowerCenter
(PWC) versions:

PWX 1 PWC 1 9.0 PWC PWC PWC PWC PWC PWC 10.0
Version 9.1.0 9.5.0 9.5.1 9.6.0 9.6.1

9.0 Yes No No No No No No

9.1.0 Yes Yes No No No No No

9.5.0 Yes Yes Yes No No No No

9.5.1 Yes Yes Yes Yes No No No

9.6.0 Yes Yes Yes Yes Yes No No

9.6.1 Yes Yes Yes Yes Yes Yes No

10.0 No No No No No No Yes

1. In this table, PWX stands for PowerExchange, and PWC stands for PowerCenter.

Hotfixes of PowerExchange are certified for use with the latest available PowerCenter hotfix.

Notes:

• If you run the PowerExchange Listener Service or PowerExchange Logger Service in the Informatica
domain, the Informatica domain must be at the same version, release, and modification (v.r.m) level as
PowerExchange.
To run different versions of the Informatica domain and PowerExchange, you must start the
PowerExchange Listener or PowerExchange Logger for Linux, UNIX, and Windows by running the dtllst or
pwxccl command-line program, respectively.

26 Chapter 2: Installing PowerExchange Client for PowerCenter (PWXPC)

CHAPTER 3

Working with Mappings

This chapter includes the following topics:

• Working with Mappings Overview, 27

• Source and Target Definitions, 28
• Working with Relational Source and Target Definitions, 28
• Working with Nonrelational Source and Target Definitions, 36
• Working with Extraction Map Definitions, 44
• Previewing PowerExchange Data in Designer, 48
• Group Source Processing in PowerExchange, 51
• Multiple-Record Writes to Nonrelational Targets, 53
• Working with Source Qualifiers, 56
• Lookup Transformations, 57
• Stored Procedure Transformations, 59

Working with Mappings Overview

A mapping is a set of source and target definitions linked by transformation objects that define the rules for
data transformation. Mappings represent the data flow between sources and targets.

Source and target definitions represent metadata for sources and targets. When you create a source
definition, its structure differs depending on the type of source it represents. The source qualifier for a source
definition also differs in structure depending on type of source definition.

After you create a source or target definition, you can include it in a mapping to extract data from the source
or load data to the target. You can extract source data in batch, change, or real-time mode.

With CDC mappings, it is generally necessary to have multiple mappings: a batch mapping to materialize the
target tables from the source tables in preparation for CDC and the CDC mapping itself which then uses
extraction map sources for the source tables. To minimize the effort in creating these mappings, create any
business rules applicable to both the batch and CDC sessions in mapplets.

27
Source and Target Definitions
To import relational and nonrelational sources from PowerExchange, in the Source Analyzer, click
Sources > Import from PowerExchange.

To import relational and nonrelational targets from PowerExchange, in the Target Designer, click
Targets > Import from PowerExchange.

In each case, the Import from PowerExchange dialog box appears. Certain fields in this dialog box appear
for sources only, not for targets.

Additional input fields appear, based on the Source Type that you choose. After you create a source or
target definition, you can edit it.

Working with Relational Source and Target

Definitions
With PowerExchange Client for PowerCenter, you can create source definitions to read from, and target
definitions to write to, the following relational databases:

• DB2 for z/OS

• DB2 for i5/OS
• DB2 for Linux, UNIX, and Windows
• Microsoft SQL Server
• Oracle

Working with DB2 Definitions

You can create a DB2 source definition in the following ways:

• Import the table definitions that PowerExchange retrieves from the DB2 catalog.
• Import DB2 or DB2 unload (DB2UNLD) data map definitions from PowerExchange.
• Import extraction map definitions for PowerExchange.
• Manually create a DB2 definition.
You can create a DB2 target definition in the following ways:

• Import the table definitions that PowerExchange retrieves from the DB2 catalog.
• Manually create a DB2 definition.
• Create a DB2 target definition from a DB2 source definition. In the Target Designer, drag a DB2 source
definition to the workspace.
Note: If the source includes a LOB column, the source definition and source qualifier include a port for the
ROWID column. The mapping does not use this port, and you can leave it unconnected to the target.
Tip: If your repository already contains DB2 definitions, you can use them to extract data from or load
data to a DB2 table. However, the metadata definition must match the table structure of the DB2 table.

28 Chapter 3: Working with Mappings

Related Topics:
• “Importing Nonrelational Source Definitions” on page 36
• “Working with Extraction Map Definitions” on page 44

Importing a DB2 Relational Source or Target Definition

You can connect to a DB2 database through a PowerExchange Listener or locally without a Listener to import
DB2 metadata for creating PowerCenter source or target definitions. This procedure describes the import
process for DB2 for z/OS, DB2 for i5/OS, or DB2 for Linux, UNIX, and Windows databases.

After you connect to the DB2 database, the Designer displays the available schemas and tables. You can
filter this metadata by schema, table name, or both. Then select a table to create the source or target
definition. When you import DB2 source or target definitions through a PowerExchange Listener or locally,
the Designer imports the primary keys.

You can map DB2 tables in PowerExchange as either DB2 data maps or DB2 database unload (DB2UNLD)
data maps for DB2 for z/OS sources. Import these types of data maps for DB2 sources in the same manner
as nonrelational data map sources.

1. To import a DB2 source definition, click Sources > Import from PowerExchange and select a source
type of DB2zOS, DB2i5OS, or DB2LUW.
To import a DB2 target definition, click Targets > Import from PowerExchange and select a source
type of DB2zOS, DB2i5OS, or DB2LUW.
2. Enter connection information.
The following table describes the required and optional connection attributes:

Attribute Required Description

or Optional

Location Required Name of the PowerExchange Listener (Node name from PowerExchange
dbmover.cfg) on the system on which the database resides.

User Name Required A user name that has the authority to connect to the database.
For connections to supported Linux, UNIX, or Windows platforms, if you
have enabled PowerExchange LDAP user authentication and disabled
relational pass-through authentication, the user name is the enterprise
user name. For more information, see the PowerExchange Reference
Manual.

Working with Relational Source and Target Definitions 29

 Attribute Required Description
or Optional

Password Required A password for the specified user. Passwords that are less than nine
characters in length are not validated.
To avoid errors that might arise due to code page differences,
Informatica recommends that the password includes only those
characters that are allowed for passphrases.
For a DB2 for i5/OS or DB2 for z/OS source or target, you can enter a
valid PowerExchange passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length. A z/OS passphrase
can be from 9 to 128 characters in length. A passphrase can contain the
following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation
marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not
affect the allowable characters in PowerExchange passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in
length. PowerExchange truncates passphrases longer than 100
characters when passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with
a security setting of SECURITY=(1,N) or higher in the DBMOVER
member. For more information, see "SECURITY Statement" in the
PowerExchange Reference Manual.

Multi-Record - Select to list multi-record data maps. SEQ or VSAM sources only.
Datamaps

Source Type Required DB2zOS, DB2i5OS, or DB2LUW.

CDC - Select to list CDC extraction maps - source only.

Datamaps

Network Optional Number of seconds that elapse during a network send or receive
Timeout operation before PowerExchange ends the connection and issues a
timeout error message.

Subsystem Id Required Name of the subsystem for connection. DB2zOS only.

Database Required Name of the database for connection. DB2i5OS and DB2LUW only.
Name

Schema Optional Enter a schema name to filter the resulting data maps.

30 Chapter 3: Working with Mappings

 Attribute Required Description
or Optional

Map name Optional Enter a table name to filter the resulting data maps.

Respect Case Optional If selected, PowerExchange filters metadata by using the exact case of
the Schema and Map name fields. Otherwise, PowerExchange ignores
the case when searching for data maps.

3. Optionally, define a filter criteria to narrow the list schemas and tables that are displayed.
To define a filter, enter a schema name, table name, or both. You can enter name patterns by using the
following wildcard characters:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
For example:
• A* displays schemas and tables that begin with an A.
• *A displays schemas and tables that end with an A.
• *cust* displays schemas that contain “cust.”
4. Click Connect.
The Designer displays the metadata to import in the Selected Datamaps box. If no tables are found, the
Designer displays “No Data Found."
5. Select the table or tables that you want to import.
To select multiple tables, use the Shift key or Ctrl key or click Select all.
6. Click OK.
The source or target definition appears in the workspace.

Related Topics:
• “Importing Nonrelational Source Definitions” on page 36

Working with Microsoft SQL Server Definitions

You can create a Microsoft SQL Server source definition in the following ways:

• In PowerExchange, import table definitions from Microsoft SQL Server.

• Import extraction map definitions from PowerExchange.
• In the PowerCenter ODBC interface, import table definitions from Microsoft SQL Server.
• Manually create a Microsoft SQL definition.
You can create a Microsoft SQL Server target definition in the following ways:

• In PowerExchange, import table definitions from Microsoft SQL Server.

• In the PowerCenter ODBC interface, import table definitions from Microsoft SQL Server.
• Manually create a Microsoft SQL definition.
Tip: If your repository contains Microsoft SQL Server definitions, you can use them to extract data from or
load data to a Microsoft SQL Server table. However, the metadata definition must match the table
structure of the Microsoft SQL Server table.

Working with Relational Source and Target Definitions 31

 Related Topics:
• “Working with Extraction Map Definitions” on page 44

Importing a Microsoft SQL Server Relational Source or Target Definition

You can connect to a Microsoft SQL Server database through a PowerExchange Listener, or locally without a
Listener, to import SQL Server metadata. When you connect to the database, the Designer displays database
schemas and tables. Before you connect to the database, you can filter the metadata the Designer displays
by schema, table name, or by both values.

Select a table to create the source or target definition. When you import a Microsoft SQL Server definition
through a PowerExchange Listener or locally, the Designer imports the primary keys.

To import a Microsoft SQL Server relational source or target definition:

1. To import a Microsoft SQL Server source definition, click Sources > Import from PowerExchange and
select a source type of MSSQL.
To import a Microsoft SQL Server target definition, click Targets > Import from PowerExchange and
select a source type of MSSQL.
2. Enter connection information.
The following table describes required and optional connection information:

Attribute Required/ Description

Optional

Location Required Name of the PowerExchange Listener (Node name from

PowerExchange dbmover.cfg) on the system on which the database
resides.

User Name Required User ID for connection to the database.

Password Required Password associated with the User ID.

Multi-Record - Select to list multi-record data maps. SEQ or VSAM sources only.
Datamaps

Source Type Required MSSQL.

CDC Datamaps - Select to list CDC extraction maps - source only

Network Timeout Optional Number of seconds that elapse during a network send or receive
operation before PowerExchange ends the connection and issues a
timeout error message.

Server Name Required SQL Server instance name.

Database Name Required Database name in the SQL instance specified.

Schema Optional Enter a schema name to filter the resulting data maps.

32 Chapter 3: Working with Mappings

 Attribute Required/ Description
Optional

Map name Optional Enter a table name to filter the resulting data maps.

Respect Case Optional If selected, PowerExchange filters data maps by using the exact
case of the Schema and Map name fields. Otherwise,
PowerExchange ignores the case when searching for data maps.

3. Optionally, enter a filter to view particular schemas and tables from the database.
Enter a schema name, table name, or both values.
You can enter a filter condition to display schemas and tables that meet the filter condition. Use one of
the following wildcard characters in the filter condition:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
Use the following syntax when entering a wildcard character:
• Enter the filter condition as a prefix. For example, enter A* to display schemas and tables that
begin with an A.
• Enter the filter condition as a suffix. For example, enter *A to display schemas and tables that end
with an A.
• Enter the filter condition as a substring. For example, enter *cust* to display schemas and tables
that contain “cust.”
4. Click Connect.
The Designer displays metadata to import in the Selected Datamaps box. If no tables are found, “No
Data Found” is displayed.
5. Select the table you want to import.
• Hold down the Shift key to select blocks of tables.
• Hold down the Ctrl key to make non-contiguous selections within a schema.
• Use the Select all button to select all tables.
• Use the Select none button to clear all highlighted selections.
6. Click OK.
The source or target definition appears in the workspace.

Working with Oracle Definitions

You can create an Oracle source definition in the following ways:

• In PowerExchange, import table definitions from Oracle.

• Import extraction map definitions from PowerExchange.
• In the PowerCenter ODBC interface, import table definitions from Oracle.
• Manually create an Oracle source definition.
You can create an Oracle target definition in the following ways:

• In PowerExchange, import table definitions from Oracle.

• In the PowerCenter ODBC interface, import table definitions from Oracle.

Working with Relational Source and Target Definitions 33

 • Manually create an Oracle source definition.
Tip: If your repository contains Oracle definitions, you can use them to extract data from or load data to
an Oracle table. However, the metadata definition must match the table structure of the Oracle table.

Related Topics:
• “Working with Extraction Map Definitions” on page 44

Importing an Oracle Relational Source or Target Definition

You can connect to an Oracle database through a PowerExchange Listener, or locally without a Listener, to
import Oracle metadata. When you connect to the database, the Designer displays database schemas and
tables. Before you connect to the database, you can filter the metadata the Designer displays by schema,
table name, or both values.

Select a table to create the source or target definition. When you import an Oracle source definition through a
PowerExchange Listener, the Designer imports the primary keys.

To import an Oracle relational source or target definition:

1. To import an Oracle source definition, click Sources > Import from PowerExchange and select a
source type of Oracle.
To import an Oracle target definition, click Targets > Import from PowerExchange and select a source
type of Oracle.
2. Enter connection information.
The following table describes required and optional connection information:

Attribute Required/ Description

Optional

Location Required Name of the PowerExchange Listener (Node name from PowerExchange
dbmover.cfg) on the system on which the database resides.

User Name Required User name for connecting to the database.

For sources or targets on supported Linux, UNIX, or Windows systems, if
you have enabled PowerExchange LDAP user authentication and
disabled relational pass-through authentication, the user name is the
enterprise user name. For more information, see the PowerExchange
Reference Manual.

Password Required Password associated with the user name.

Multi-Record - Select to list multi-record data maps. SEQ or VSAM sources only.
Datamaps

Source Type Required ORACLE.

CDC Datamaps - Select to list CDC extraction maps - source only.

Network Timeout Optional Number of seconds that elapse during a network send or receive
operation before PowerExchange ends the connection and issues a
timeout error message.

34 Chapter 3: Working with Mappings

 Attribute Required/ Description
Optional

TNS Name Required Oracle SID (instance name).

Schema Optional Enter a schema name to filter the resulting data maps.

Map name Optional Enter a table name to filter the resulting data maps.

3. Optionally, enter a filter to view particular schemas and tables from the database.
Enter a schema name, table name, or both values.
You can enter a filter condition to display schemas and tables that meet the filter condition. Use one of
the following wildcard characters in the filter condition:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
Use the following syntax when entering a wildcard character:
• Enter the filter condition as a prefix. For example, enter A* to display schemas and tables that
begin with an A.
• Enter the filter condition as a suffix. For example, enter *A to display schemas and tables that end
with an A.
• Enter the filter condition as a substring. For example, enter *cust* to display schemas and tables
that contain “cust.”
4. Click Connect.
The Designer displays metadata to import in the Selected Datamaps box. If no tables are found, “No
Data Found” is displayed.
5. Select the table you want to import.
• Hold down the Shift key to select blocks of tables.
• Hold down the Ctrl key to make noncontiguous selections within a schema.
• Use the Select all button to select all tables.
• Use the Select none button to clear all highlighted selections.
6. Click OK.
The source or target definition displays in the workspace.

Editing Relational Source and Target Definitions

You can edit relational source and target definition to make minor changes to the definition. For example, you
can:

• Modify column names.

• Modify column datatypes.
• Add or delete columns.
• Modify column key relationships.
• Add a description of the definition.

Note: If you use the relational source in a CDC session, you do not need to add the DTL__CAPXACTION
field nor do you need to include an Update Strategy transformation. PWXPC will automatically include the

Working with Relational Source and Target Definitions 35

 DTL__CAPXACTION column in its SELECT statement for CDC sources. It then uses the value of the
DTL__CAPXACTION to construct the appropriate SQL statement (INSERT, UPDATE, or DELETE).

Working with Nonrelational Source and Target

Definitions
With the PowerExchange Client for PowerCenter, you can create source definitions and target definitions for
the following nonrelational source to target types:

• Adabas
• Datacom - source only
• DB2 data maps (DB2MAP) - source only
• DB2 for z/OS unload files (DB2UNLD) - source only
• IDMS - source only
• IMS
• SEQ
• VSAM

Importing Nonrelational Source Definitions

You can connect to PowerExchange locally or to a PowerExchange Listener to import a nonrelational data
maps to create source definitions or a data mapped DB2 definition.

When you connect to PowerExchange, the Designer displays data map schemas and tables for the source
type specified. Before you connect to PowerExchange, you can filter this information based on a schema
name, data map name, or both. Then select a data map to create the source definition.

The following procedure imports single- and multi-record data maps to create source definitions. You can
import multiple-record data maps for IMS unload data sets, VSAM data sets, and sequential files. For
multiple-record data maps, select the Multi-Record Datamaps attribute.

1. Click Sources > Import from PowerExchange.

2. In the Import from PowerExchange dialog box, enter information for importing a nonrelational data
map.

36 Chapter 3: Working with Mappings

The following table describes the required and optional attributes:

Attribute Required Description

or
Optional

Location Required Name of the node where the PowerExchange Listener runs, as specified in
a NODE statement in the PowerExchange DBMOVER configuration file that
is on the system where the database resides.

User Name Required A user name that has the authority to connect to the database.
For sources on supported Linux, UNIX, or Windows systems, if you have
enabled PowerExchange LDAP user authentication, the user name is the
enterprise user name. For more information, see the PowerExchange
Reference Manual.

Password Required A password for the specified user. Passwords that are less than nine
characters in length are not validated.
To avoid errors that might arise due to code page differences, Informatica
recommends that the password includes only those characters that are
allowed for passphrases.
For nonrelational sources on z/OS, and for sequential source files on i5/OS
for bulk data movement, you can enter a valid PowerExchange passphrase
instead of a password. An i5/OS passphrase can be from 9 to 31
characters in length. A z/OS passphrase can be from 9 to 128 characters in
length. A passphrase can contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation
marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not affect
the allowable characters in PowerExchange passphrases.
Note: On z/OS, valid RACF passphrase can be up to 100 characters in
length. PWXPC truncates passphrases longer than 100 characters when
passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member.
For more information, see "SECURITY Statement" in the PowerExchange
Reference Manual.

Multi-Record Optional Select this option to list multiple-record data maps, which are available for
Datamaps IMS, SEQ, and VSAM sources only. For more information, see “ Group
Source Processing in Bulk Data Movement Sessions with Nonrelational Data
Sources” on page 51 .

Use Optional Select this option to generate sequence fields. IMS unload sources only.
Sequence You can select this option only if you also select Multi-Record Data Maps.
Fields For more information, see “Multiple-Record Writes to Nonrelational Targets” on
page 53 .

Working with Nonrelational Source and Target Definitions 37

 Attribute Required Description
or
Optional

Source Type Required Select one of the following source types: ADABAS, DATACOM, DB2MAP,
DB2UNLD, IDMS, IMS, SEQ, or VSAM.

CDC - Select this option to list CDC extraction maps - source only.
Datamaps

Network Optional The number of seconds that elapse during a network send or receive
Timeout operation before PowerExchange ends the connection and issues a timeout
error message.

Schema Optional A schema name to use for filtering the data maps.

Map name Optional A data map name to use for filtering the data maps.

Respect Optional Select this option to have PowerExchange use case-sensitive filtering of
Case the data maps based on the values in the Schema and Map name fields.
When this option is not selected, PowerExchange ignores case when
searching for data maps.

Selected - Lists the available data maps for the connection, source type, and filter
Datamaps details that you entered.

3. Optionally, define a filter to narrow the list of schemas and tables.

To define a filter condition, enter a schema name, a table name, or both. You can enter name patterns
by using the following wildcard characters:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
For example:
• A* displays schemas and tables that begin with an A.
• *A displays schemas and tables that end with an A.
• *cust* displays schemas that contain “cust.”
4. Click Connect.
The Designer displays the metatdata for the data maps in the Selected Datamaps box. If no data maps
are found, the Designer displays “No Data Found."
5. Select the data map or data maps that you want to import.
To make multiple selections, use the Shift key or Ctrl key or click Select all.
6. Click OK.
The source definitions are displayed. The Designer uses the data map names as the names of the
source definition.

38 Chapter 3: Working with Mappings

Importing Nonrelational Target Definitions
You can connect to PowerExchange locally or to a PowerExchange Listener to import a nonrelational target
definition.

When you connect to PowerExchange, the Designer displays data map schemas and tables for the specified
target type. Before you connect to PowerExchange, you can filter the metadata by schema, data map name,
or both. Then select a data map to create the target definition.

1. To import a nonrelational target definition, click Targets > Import from PowerExchange and select the
target type.
The dialog box and parameters displayed are the same for each nonrelational target type.
2. Enter connection information.
The following table describes required and optional connection attributes:

Attribute Required Description

or Optional

Location Required Name of the PowerExchange Listener (Node name from PowerExchange
dbmover.cfg) on the system on which the database resides.

User Name Required A user name that has the authority to connect to the database.
For targets on supported Linux, UNIX, or Windows systems, if you have
enabled PowerExchange LDAP user authentication, the user name is the
enterprise user name. For more information, see the PowerExchange
Reference Manual.

Working with Nonrelational Source and Target Definitions 39

 Attribute Required Description
or Optional

Password Required A password for the specified user. Passwords that are less than nine
characters in length are not validated.
To avoid errors that might arise due to code page differences, Informatica
recommends that the password includes only those characters that are
allowed for passphrases.
For nonrelational targets on z/OS, and for sequential target files on i5/OS
for bulk data movement, you can enter a valid PowerExchange passphrase
instead of a password. An i5/OS passphrase can be from 9 to 31
characters in length. A z/OS passphrase can be from 9 to 128 characters
in length. A passphrase can contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation
marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not affect
the allowable characters in PowerExchange passphrases.
Note: On z/Os, valid RACF passphrase can be up to 100 characters in
length. PWXPC truncates passphrases longer than 100 characters when
passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member.
For more information, see "SECURITY Statement" in the PowerExchange
Reference Manual.

Multi-Record Optional Select to list multi-record data maps. IMS unload targets only. For more
Datamaps information, see “Multiple-Record Writes to Nonrelational Targets” on page
53 .

Use Optional Select to generate sequence fields in the target. IMS unload targets only.
Sequence You can select this option only if you also select Multi-Record Data Maps.
Fields For more information, see “Multiple-Record Writes to Nonrelational
Targets” on page 53 .

Source Type Required Select one of the following target types: ADABAS, IMS, SEQ, or VSAM.

Network Optional Number of seconds that elapse during a network send or receive operation
Timeout before PowerExchange ends the connection and issues a timeout error
message.

Schema Optional Enter a schema name to filter the resulting data maps.

Map name Optional Enter a data map name to filter the resulting data maps.

40 Chapter 3: Working with Mappings

 Attribute Required Description
or Optional

Respect Optional If selected, PowerExchange filters data maps by using the exact case of
Case the Schema and Map name fields. Otherwise, PowerExchange ignores the
case when searching for data maps.

Selected n/a Lists the available data maps for the connection, database and filter details
Datamaps that you entered.

3. Optionally, define a filter to narrow the list of schemas and tables.

To define a filter condition, enter a schema name, a table name, or both. You can enter name patterns
by using the following wildcard characters:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
For example:
• A* displays schemas and tables that begin with an A.
• *A displays schemas and tables that end with an A.
• *cust* displays schemas that contain “cust.”
4. Click Connect.
The Designer displays the metadata for the data maps to import in the Selected Datamaps box. If no
data maps are found, the Designer displays “No Data Found.”
5. Select the data map or data maps that you want to import.
To make multiple selections, use the Shift key or Ctrl key or click Select all.
6. Click OK.
The target definitions are displayed. The Designer uses the data map names as the names of the target
definition.

Viewing Nonrelational Source and Target Definition Details

Nonrelational source and target definitions contain information about the data map structure on the Attributes
tab. Additional metadata information is included on the Metadata Extensions tab.

Viewing Data Map Information

The Attributes tab of a nonrelational source or target definition lists each field in the definition. It also lists the
attributes of each field and their values from the data map used to create the definition. The attributes include
information such as the name of the base record to which the field belongs.

The following table lists the attributes the Attributes tab displays for each field in a nonrelational definition:

Attribute Name Description

column_name Name of the field in the data map.

base_rec Name of the record to which the field belongs. This corresponds to name of the group the
field belongs to in the source definition.

Working with Nonrelational Source and Target Definitions 41

 Attribute Name Description

base_fld Name of the base record and table field name in the following format:
<Base_Field_Name>:<Table_Field_Name>

base_fld_tpe PowerExchange datatype of the field.

base_fld_offset Offset value from which the field starts. For example, if the value is 5, the field starts at the
fifth position. You determine the offset value of each field based on the order of fields in the
data map.

Viewing Metadata Extensions

Nonrelational source and target definitions have metadata extensions that displays additional information
from the data map used to create the definition. With VSAM and sequential definitions, the File Name
extension field can be modified.

You can override the data map name and schema name in the session properties.

The following table describes the extensions on the Metadata Extensions tab for a nonrelational definition:

Extension Name Description

Access Method Method that you specified in the data map to access the source database:
- A = Adabas
- D = IMS DL1
- E = VSAM ESDS
- I = IDMS
- K = VSAM KSDS
- O = IMS ODBA
- N = VSAM RRDS
- S = sequential (SEQ)
- W = DB2 unload file (DB2UNLD)
- X = Datacom
- Z = DB2 data map

comments Any comments.

File Name Name of the data set (sequential or VSAM only).

Hierarchy Structure For source and target definitions for which you selected Use Sequence Fields, the
XML that describes the hierarchy structure of the data. Click the arrow in the Value
column to display the XML.
For more information, see “Hierarchy Structure” on page 55.

Map Name Name of the data map.

Schema Name Name of the schema on which you based the data map.

Use Sequence Fields Whether to generate sequence fields for the source and target definitions. For more
information, see “Multiple-Record Writes to Nonrelational Targets” on page 53.

42 Chapter 3: Working with Mappings

 Related Topics:
• “Editing Metadata Extensions ” on page 43
• “Configuring Batch Mode Sessions for Nonrelational Sources” on page 138

Editing Nonrelational Source and Target Definitions

You can edit nonrelational source and target definitions to make minor changes to the definition. For
example, you can:

• Modify a column datatype

• Modify the owner name
• Modify column key relationships
• Add or delete columns
• Add a description of the definition
• Create metadata extensions

Note: If you use the nonrelational source in a CDC session you do not need to add the DTL__CAPXACTION
field nor do you need to include an Update Strategy transformation. PWXPC will automatically include the
DTL__CAPXACTION column in its SELECT statement for CDC sources. It then uses the value of the
DTL__CAPXACTION to construct the appropriate SQL statement (INSERT, UPDATE, or DELETE).

Editing Metadata Extensions

With VSAM and sequential definitions, the File Name extension field can be modified. All other field
populated by PWXPC when a data map is imported cannot be changed.

By default, the File Name field contains the value from the PowerExchange data map File Name field which
contains the VSAM data set name or the sequential file name.

Note: Any VSAM and sequential definitions created prior to PowerCenter 8.5 will have a blank File Name. If
desired, this field can be populated manually or by re-importing the data map.

The following procedure explains how to manually update the File Name field to add or change the file name
in a VSAM or sequential definition.

To edit metadata extensions:

1. Double-click the source or target definition in the workspace.

2. Select the Metadata Extensions tab.
PWXPC populates the File Name field with the File Name for the PowerExchange data map.
3. Click in the Value field and select the down arrow.
The Edit Metadata Extension Value panel appears, allowing you to enter or change the file name.
4. Click OK.
5. Click OK to end Edit Tables.
You can also re-import the PowerExchange data map to automatically populate the File Name field with the
information contained in the data map.

Working with Nonrelational Source and Target Definitions 43

Working with Extraction Map Definitions
When creating mappings for change data capture (CDC), import the extraction map rather than the source
table metadata.

Extraction maps provide the following benefits:

• Source extraction maps can contain a subset of the source columns to reduce capture processing.
• Extraction maps contain the PowerExchange-generated columns, such as DTL__CAPXTIMESTAMP and
DTL__CAPXACTION, and DTL__BI, which provided information needed for CDC.
To import an extraction map, connect to source location where the extraction maps reside, either locally or
through a PowerExchange Listener.

After you connect to PowerExchange, the Designer displays the schemas and tables in the extraction map.
You can filter the extraction maps by schema name, map name, or by both. Then select an extraction map to
create the source definition.

1. To import an extraction map for a CDC source definition, click Sources > Import from PowerExchange
and select CDC Datamaps. Then select the source type.
The dialog box and parameters displayed are the same for each nonrelational source type.
2. Enter connection information.
The following table describes required and optional connection attributes:

Attribute Required Description

or Optional

Location Required Node name of the PowerExchange Listener, as specified in a NODE

statement in the PowerExchange DBMOVER configuration file on the
system where the database resides.

User Name Required A user name that has the authority to connect to the database.
For sources on supported Linux, UNIX, or Windows systems, if you have
enabled PowerExchange LDAP user authentication, the user name is the
enterprise user name. For more information, see the PowerExchange
Reference Manual.

44 Chapter 3: Working with Mappings

Attribute Required Description
or Optional

Password Required A password for the specified user. Passwords that are less than nine
characters in length are not validated.
To avoid errors that might arise due to code page differences, Informatica
recommends that the password includes only those characters that are
allowed for passphrases.
For access to i5/OS or z/OS, you can enter a valid PowerExchange
passphrase instead of a password. An i5/OS passphrase can be from 9 to
31 characters in length. A z/OS passphrase can be from 9 to 128
characters in length when using PWXPC. A passphrase can contain the
following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation
marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not affect
the allowable characters in PowerExchange passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in
length. PWXPC truncates passphrases longer than 100 characters when
passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member.
For more information, see "SECURITY Statement" in the PowerExchange
Reference Manual.

Multi-Record - Select this option to list multi-record data maps. SEQ or VSAM sources
Datamaps only.

Source Type Required One of the following source types:

- ADABAS
- DATACOM
- DB2zOS
- DB2i5OS
- DB2LUW
- IDMS
- IMS
- MSSQL
- ORACLE
- VSAM

CDC Required Select this option to list CDC extraction maps.

Datamaps

Network Optional The number of seconds that must elapse during a network send or receive
Timeout operation before PowerExchange ends the connection and issues a
timeout error message.

Schema Optional A schema name that is used to filter the resulting data maps.

Working with Extraction Map Definitions 45

 Attribute Required Description
or Optional

Map name Optional Enter a data map name to filter the data maps.

Respect Optional Select this option to have PowerExchange filter data maps by using the
Case exact case of the values in the Schema and Map name fields. When this
option is not selected, PowerExchange ignores the case when searching
for data maps.

3. Optionally, define a filter criteria to narrow the list schemas and tables that are displayed.
To define a filter condition, enter a schema name, a table name, or both. You can enter name patterns
by using the following wildcard characters:
• * (asterisk). Represents one or more characters.
• ? (question mark). Represents one character.
For example:
• A* displays schemas and tables that begin with an A.
• *A displays schemas and tables that end with an A.
• *cust* displays schemas that contain “cust.”
4. Click Connect.
The Designer displays metadata to import in the Selected Datamaps box. If no tables are found, “No
Data Found” displays. The extraction maps shown will be filtered based on the source type specified.
5. Select the extraction map or maps that you want to import.
To select multiple extraction maps, use the Shift key or Ctrl key or click Select all.
6. Click OK.
The source definition appears in the workspace. The Designer uses the extraction map name as the
name of the source definition.

Viewing Extraction Map Definition Details

Extraction map definitions contain information about the data map structure on the Attributes tab. They also
contain metadata extensions on the Metadata Extensions tab.

Viewing Extraction Map Information

The Attributes tab of an extraction map definition lists each field in the source definition. It also lists the
attributes of each field and their values based on the extraction map from which you imported the source
definition. The attributes include information such as the name of the base record to which the field belongs.

The following table describes the attributes that the Attributes tab displays for each field in a source
definition:

Attribute Name Description

column_name Name of the field in the extraction map.

base_rec blank.

46 Chapter 3: Working with Mappings

 Attribute Name Description

base_fld blank

base_fld_tpe blank.

base_fld_offset blank

Viewing Metadata Extensions in Extraction Maps

Extraction map definitions have metadata extensions that display information about the data map on which
you based the source definition. You cannot modify these metadata extensions. However, you can override
the data map name and schema name in the session properties.

The following table describes the extensions on the Metadata Extensions tab for an extraction map:

Extension Name Description

Access Method Method you specified in the data map to access the source database:
- A = Adabas
- B = DB2 for z/OS and DB2 for i5/OS
- D = IMS
- E = VSAM ESDS
- I = IDMS
- K = VSAM KSDS
- L = MSSQL
- N = VSAM RRDS
- P = Oracle
- V = DB2 for Linux, UNIX, and Windows
- X = Datacom

comments Any comments.

Map Name Name of the extraction map.

Original Name Original table name in relational database or PowerExchange data map.

Original Schema Original schema or owner name in relational database or PowerExchange data map.

Schema Name Name of the schema for the extraction map.

Related Topics:
• “Configuring Batch Mode Sessions for Nonrelational Sources” on page 138

Editing an Extraction Map Definition

You can edit extraction maps to make minor changes to them, for example:

• Modify a column datatype.

• Modify the owner name.

Working with Extraction Map Definitions 47

 • Add or delete columns.
• Add a description of the extraction map.
• Create metadata extensions.
When you use extraction maps, you do not need to add the DTL__CAPXACTION field nor do you need to
include an Update Strategy transformation. PWXPC will automatically include the DTL__CAPXACTION
column in its SELECT statement for CDC sources. It then uses the value of the DTL__CAPXACTION to
construct the appropriate SQL statement (INSERT, UPDATE, or DELETE).

Warning: Changing column information in the extraction map could result in failures in the session or
workflow during the extraction process.

Previewing PowerExchange Data in Designer

You can preview data in the Designer for any valid relational and nonrelational source or target definitions.
You can also preview change data for extraction map source definitions. You cannot preview data for multi-
record data map source definitions.

Previewing data allows you to see the source or target data. The data that is shown depends upon the
database type of the source or target definition. There are three database types for PowerExchange source
and target definitions:

• Relational, such as DB2 for DB2 metadata.

• Nonrelational, such as PWX_VSAM_NRDB2 for VSAM data maps.
• Extraction Maps, such as PWX_DB2390_CDC for DB2 CDC data maps.
You can preview source or target data in the following Designer tools:

• Source Analyzer. Preview source data in the Source Analyzer after you import the source.
• Target Designer. Preview target data in the Target Designer after you import a target.
• Mapplet Designer. Preview source data in the Mapplet Designer while creating a mapplet.
• Mapping Designer. Preview source and target data in the Mapping Designer while creating a mapping.

Previewing PowerExchange Relational Source or Target Data

Preview data for relational source or target tables based on the source or target relational definitions.

Note: You must be able to connect to PowerExchange on the source or target system. To preview relational
data, you must use an ODBC data source to connect to PowerExchange.

1. Select a relational source or target definition in the workspace.

2. Right-click the source or target definition in the workspace and click Preview Data.
The Preview Data dialog box appears.
3. Select an ODBC data source name.
You can add an ODBC data source by clicking the Browse button.
4. If necessary, enter Username and Password values.

48 Chapter 3: Working with Mappings

 For PowerExchange sources and targets on i5/OS or z/OS, a user name and password are required only
if connecting to a PowerExchange Listener that is configured with a PowerExchange security setting of
SECURITY=(1,x) or (2,x) in the DBMOVER configuration file.
To connect to sources or targets on i5/Os or z/OS, you can enter a valid PowerExchange passphrase
instead of a password. An i5/OS passphrase can be from 9 to 31 characters in length. A z/OS
passphrase can be from 9 to 79 characters in length when using ODBC. A passphrase can contain
uppercase and lowercase letters, numbers, spaces, and the following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
5. Enter a table owner name.
6. Enter the number of rows that you want to preview.
Default is 100. The Preview Data dialog box can display up to 500 rows and up to 65,000 columns.
7. Click Connect.
The table contents appears in the Preview Data dialog box.
8. To change the number of rows that are displayed, enter another number of rows and click Refresh.
9. When you are done previewing data, click Close.

Previewing PowerExchange Nonrelational Source or Target Data

Preview data from nonrelational sources or targets based on the source or target definitions. For DB2, the
preview process displays data from the source DB2 table.

To preview data, you must be able to connect to PowerExchange on the source or target system.
PowerCenter uses PWXPC to connect to PowerExchange for nonrelational sources or targets.

1. Select a nonrelational source or target definition in the workspace.

2. Right-click the source or target definition in the workspace and click Preview Data.
The Preview Data dialog box appears.
3. Select a Location name.
The location names are retrieved from the NODE statements in the dbmover.cfg file on the Designer
system. To add locations, add NODE statements to the dbmover.cfg file.
4. If necessary, enter Username and Password values.
For PowerExchange sources and targets on i5/OS and z/OS, a user name and password is required only
if connecting to a PowerExchange Listener that is configured with a PowerExchange security setting of
SECURITY=(1,x) or (2,x) in the DBMOVER configuration file.
To connect to sources or targets on z/OS, or to connect to sequential source or target files on i5/OS for
bulk data movement, you can enter a valid PowerExchange passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length. A z/OS passphrase can be from 9 to 128
characters in length when using a PWXPC connection. A passphrase can contain uppercase and
lowercase letters, numbers, spaces, and the following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
The first character is an apostrophe.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in length. PowerExchange
truncates passphrases longer than 100 characters when passing them to RACF for validation.
5. In the Schema field, change the schema name from the data map, if necessary.

Previewing PowerExchange Data in Designer 49

 6. In the Network Timeout field, enter the number of seconds that must elapse during a network send or
receive operation before PowerExchange ends the connection and issues a timeout error message.
Default is 180.
7. Enter the number of rows that you want to preview.
Default is 10. The Preview Data dialog box can display up to 500 rows and up to 65,000 columns.
8. Click Connect.
The source or target contents appears in the Preview Data dialog box.
9. To view more data, click More.
10. To change the number of rows that are displayed, enter another number of rows and click Refresh.
11. When you are done previewing data, click Close.

Previewing PowerExchange Change Data

Preview data based on extraction map definitions (CDC Datamaps) to display change data from the change
stream.

You must be able to connect to PowerExchange on the source platform to preview data. PowerCenter uses
PWXPC to connect to PowerExchange for previewing CDC data based on extraction maps.

1. Select an extraction map source definition in the workspace.

2. Right-click the source definition in the workspace and click Preview Data.
The Preview Data dialog box appears.
3. Select a Location name.
The location names are retrieved from the NODE statements in the dbmover.cfg configuration file on the
Designer platform. To add additional locations, add NODE statements to the dbmover.cfg file.
4. If necessary, enter Username and Password values.
For PowerExchange sources and targets on z/OS, a user name and password are required only if you
are connecting to a PowerExchange Listener that is configured with a PowerExchange security setting of
SECURITY=(1,x) or (2,x) in the DBMOVER configuration file.
To connect to sources on i5/OS or z/OS, you can enter a valid PowerExchange passphrase instead of a
password. An i5/OS passphrase can be from 9 to 31 characters in length. A z/OS passphrase can be
from 9 to 128 characters in length when using a PWXPC connection. A passphrase can contain
uppercase and lowercase letters, numbers, spaces, and the following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
The first character is an apostrophe.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in length. PowerExchange
truncates z/OS passphrases longer than 100 characters when passing them to RACF for validation.
5. In the Schema field, change the schema name from the extraction map, if necessary.
6. In the Network Timeout field, enter the number of seconds that must elapse during a network send or
receive operation before PowerExchange ends the connection and issues a timeout error message.
Default is 180.
7. Select Real Time or Change.
The Real Time option extracts data in real time from the change stream by using the earliest starting
restart point. The Change option extracts data from condense files by using the earliest restart point.
8. Enter the number of rows that you want to preview.
Default is 10. The Preview Data dialog box can display up to 500 rows and up to 65,000 columns.

50 Chapter 3: Working with Mappings

 9. Click Connect.
The source contents appears in the Preview Data dialog box.
10. To return more data, click More.
When the end of log is reached, the More button becomes unavailable.
11. To change the number of rows to preview, enter another number of rows and click Refresh.
12. When you done previewing data, click Close.

Related Topics:
• “Default Restart Points” on page 169

Group Source Processing in PowerExchange

With group source processing, PowerExchange reads data from the same physical source in a single pass.
This processing enhances throughput and reduces resource consumption by eliminating multiple reads of the
source data.

PowerExchange and PWXPC use group source processing when possible in CDC sessions for all data
sources and in bulk data movement sessions for nonrelational data sources.

For CDC data, PowerExchange uses group source processing to read changes for all tables defined in a
mapping from the change stream in a single pass. For nonrelational bulk data, PowerExchange uses group
source processing to read all records in a single multi-group source qualifier in a mapping.

Group Source Processing in Bulk Data Movement Sessions with

Nonrelational Data Sources
If you use PWXPC connections for bulk data movement operations, PowerExchange uses group source
processing for the following nonrelational data sources:

• IMS unload data sets

• Sequential data sets and flat files
• VSAM data sets

Source Definitions for Sources with Multiple Record Types

You create a nonrelational source definition by importing a data map from a PowerExchange Listener. For
data sources with multiple record types, the PowerExchange data map can define a record and a table for
each unique record type. The table represents a relational view of the related record.

Nonrelational source definitions present the data map metadata in groups. Each group represents a table in
the data map and contains metadata for the fields in the table.

For IMS, VSAM, and sequential or flat file data sources, use Designer to import data maps that define
multiple record types to create one or more PowerCenter source definitions. To import a multiple-record data
map, select Multi-Record Datamaps in the Import from PowerExchange dialog box.

To create a source definition for a single record, import a single table from the data map. The source
definition then contains only a single group. To create a source definition that includes all record types,

Group Source Processing in PowerExchange 51

 import the entire multiple-record data map. The source definition then contains a group for each table in the
data map. Each group contains metadata for the fields in a table.

The following figure shows a nonrelational source definition for a VSAM data map that contains multiple
tables representing multiple records in the VSAM file:

1. A section represents a table in the PowerExchange data map.

2. Group name.

In this example, the source definition contains four groups: V07A_RECORD_LAYOUT,

V07B_RECORD_LAYOUT, V07C_RECORD_LAYOUT and V07D_RECORD_LAYOUT.

Group Source Processing and Multiple-Record Data Maps

You can use group source processing with a PowerExchange multiple-record data map for a nonrelational
source with multiple record types.

When you import the data map into PowerCenter, select the Multi-Record Datamaps option in the Import
from PowerExchange dialog box. PowerCenter creates a source definition that includes all of the record
types that are defined in the data map. When a bulk data movement session with a mapping that contains the
source definition runs, PowerExchange uses group source processing to read all of the records in the source
data set or file in a single pass.

If you do not select the Multi-Record Datamaps option in the Import from PowerExchange dialog box,
PowerCenter creates a separate source definition for each table, or record, in the multiple-record data map.
When a session with a mapping that contains these single-record source definitions runs, PowerExchange
reads the source data set of file once for each source definition and does not use group source processing.

Note: PWXPC establishes a connection to the source for each source definition to read source data.

If you import IMS data maps as multiple-record data maps, you can use the source definitions only to process
IMS unload data sets. You cannot use multiple-record IMS source definitions to read all segments from an
IMS database in a single pass. For bulk data movement sessions with IMS source databases, create
mappings that contain a source definition for each segment in the IMS database.

52 Chapter 3: Working with Mappings

 Group Source Processing in CDC Sessions
When you use PWX CDC application connections to extract change data, PowerExchange uses group source
processing and reads the change stream in a single pass for all source definitions in the mapping.

When you run a CDC session, PWXPC passes a source interest list that contains all of the sources.
PowerExchange uses the source interest list to determine the sources for which to read data from the change
stream. When PowerExchange encounters changes for a source in the interest list, it passes the change data
to PWXPC. PWXPC then provides the change data to the appropriate source in the mapping.

To create source definitions in Designer that can be used to extract change data, import source metadata by
using one of the following methods:

• Import a PowerExchange extraction map by using the Import from PowerExchange dialog box.
• Import the table definitions from relational databases, by using either the Import from PowerExchange
dialog box or the Import from Database dialog box.

Restriction: To read change data for nonrelational sources, you must import extraction maps from
PowerExchange.

Informatica recommends that you use extraction maps to create source definitions for all CDC sources. When
you create source definitions from extraction maps, the mapping and session creation process is simpler for
the following reasons:

• The source definition contains the extraction map name, which eliminates the need to provide it when you
configure the session.
• The source definition contains the PowerExchange-defined CDC columns, which eliminates the need to
add these columns to the source definition. The PowerExchange-defined columns include the change
indicator and before image columns as well as the DTL__CAPX columns.

When you extract change data, PowerExchange uses group source processing for all source definitions in
the mapping. All source definitions must be for the same data source type, such as DB2, IMS, VSAM, or
Oracle. Do not include multiple data source types in the mapping. Otherwise, the session fails with message
PWXPC_10080.

For example, you cannot run a CDC session that contains a mapping with both VSAM and IMS source
definitions, even though the change stream is the same. To extract change data for both IMS and VSAM data
sources, create a mapping and session for the VSAM sources and a separate mapping and session for the
IMS sources. PowerExchange reads the change stream twice, once for the session with VSAM sources and
once for the session with IMS sources.

If you create a workflow that contains multiple CDC sessions, PowerExchange uses a connection for each
session, even if the sessions extract change data from the same change stream, such as the
PowerExchange Logger for MVS.

Multiple-Record Writes to Nonrelational Targets

During bulk data movement sessions, PowerExchange can use a multiple-record data map to read data from
a nonrelational data source and write the multiple record types in a single pass to a nonrelational target. This
process is called a multiple-record write.

When PowerExchange performs a multiple-record write, it preserves source sequencing information. To

enable multiple-record writes with sequencing, select the Multi-Record Datamaps and Use Sequence
Fields options in the Import from PowerExchange dialog box for both the source and target definition.

Multiple-Record Writes to Nonrelational Targets 53

 When you select these options, PowerExchange generates primary key and foreign key fields for each group
in the source and target definitions and also stores metadata about these record relationships in XML format.

You can perform multiple-record writes with the following types of sources and targets:

• Source and target IMS unload data sets on z/OS

• Any combination of the following types of nonrelational sources and targets:
- Sequential or VSAM ESDS, KSDS, or RRDS data sets on z/OS

- Files on Linux, UNIX, or Windows

- Files on i5/OS

Related Topics:
• “Considerations for Sessions that Perform Multiple-Record Writes” on page 135
• “Connection Attributes for Multiple-Record Writes” on page 129

Group Processing with Sequencing

To perform multiple-record writes with sequencing, PowerExchange uses group source processing to read
source data in a single pass and uses group target processing to write data to the target in a single pass.

PowerExchange generates sequence fields to pass metadata about the source record relationships to the
target. After you enable multiple-record writes with sequencing for a PowerCenter workflow, the workflow can
read the multiple-record source, use the generated sequence fields to preserve the sequencing information,
and write data to the target in the same record sequence as in the source.

To determine the relationships among IMS record types, PowerExchange uses the segment information that
you imported from an IMS DBD into the multiple-record data map. To determine the relationships between
record types for sequential or VSAM sources, PowerExchange uses the record properties that you define in
the data map for the source or target. These record properties include the parent record name, the name of
the record that the current record redefines, and whether the record is a header or trailer. PowerExchange
uses the segment information or record properties to generate primary and foreign keys and sequencing
metadata in the source and target definitions.

When you run a PowerCenter session, PowerExchange uses the generated key values to reconstruct and
write the data to the target in the correct sequence. PowerExchange maintains the data in a sequencing and
queuing cache on the Integration Service node. When PowerExchange writes data to the target, it deletes the
generated key fields and sends the data across the network in the correct sequence to the target file.

For group source processing with sequencing, PowerExchange performs the following functions when it
reads the source data:

• Reads the data in a single pass

• Drops any complex tables in the data map from the source definition
• Adds primary-key and foreign-key columns to the source definition
• Generates sequencing metadata to represent the source hierarchy
For group target processing with sequencing, PowerExchange performs the following functions when it writes
the target data:

• Resequences the records in the cache.

• Discards sequence data fields after sequencing the records and before writing them to the target.
• Writes the data to the target in a single pass.

54 Chapter 3: Working with Mappings

Sequencing Support in Source and Target Definitions
If you enable multiple-record write operations with sequencing when you import a multiple-record data map
into PowerCenter, the resulting source or target definition includes special features to support record
sequencing.

The following PowerCenter features support sequencing:

• The generated primary key and foreign key columns in the table definitions. The keys contain sequence
numbers that uniquely identify each row and its parent.
• The following metadata extensions on the Metadata Extensions tab of the Edit Tables dialog box:
- Use Sequence Fields. This metadata extension indicates that you selected the Use Sequence Fields
option in the Import from PowerExchange dialog box.
- Hierarchy Structure. This metadata extension is a segment of XML code that defines the hierarchy
structure of the tables in a source or target definition.

Note: The Database type field on the Table tab of the Edit Tables dialog box contains different values for
sequential source and target definitions. The database type for sequential source definitions is PWX_VSAM.
The database type for sequential target definitions is PWX_SEQ.

Generated Primary Key and Foreign Key Columns

When you enable sequencing in a source definition, the source definition includes generated primary key and
foreign key columns. These key columns are used to capture sequencing information from the source file and
reconstruct the data in the same sequence in the target.

The source definition includes the following generated columns:

• DTL_PK_tablename. Contains a sequence number that is generated for each source table. This sequence
number serves as the primary key.
• DTL_FK_tablename. Contains a sequence number that is generated for each table that corresponds to a
non-root record. This sequence number serves as a foreign key.
The mapping must pass the generated sequence numbers, unaltered, to the corresponding generated
columns on the target. The names and types of the generated columns must also remain unaltered. The
names of the sequence fields are contained in the XML code that is defined in the Hierarchy Structure
metadata extension.

Hierarchy Structure
The Hierarchy Structure metadata extension is a segment of XML code that defines the hierarchical structure
of the tables in a source or target definition for which you enable sequencing. Developers can use this
information to write mapping logic that controls the relationships among tables in PowerCenter
transformations.

The Hierarchy Structure includes the following information:

• Database type, data map, and schema

• Node definition for each record in the data map
Each node defines the following information:

• Node name
• Name of parent node (null string for root nodes)
• Horizontal and vertical position
• Record name

Multiple-Record Writes to Nonrelational Targets 55

 • Table name
• Column name of the generated primary key
• For non-root nodes, the column and name of the generated foreign key
The following lines of XML are an example of a Hierarchy Structure:
?<?xml version="1.0" encoding="UTF-16" standalone="no" ?>
<HIERARCHY DBTYPE="IMS" MAP="imsstudkey" SCHEMA="mrecims">

<NODE NAME="STUDENT" PARENTNODE="" POSHOR="1" POSVERT="1">

<RECORD NAME="STUDENT">
<TABLE NAME="STUDENT">
<COL_PK CANDIDATE="1" NAME="DTL_PK_STUDENT"/>
</TABLE>
</RECORD>
</NODE>

<NODE NAME="COURSE" PARENTNODE="STUDENT" POSHOR="1" POSVERT="2">

<RECORD NAME="COURSE">
<TABLE NAME="COURSE">
<COL_PK CANDIDATE="1" NAME="DTL_PK_COURSE"/>
<COL_FK CANDIDATE="1" NAME="DTL_FK_COURSE"/>
</TABLE>
</RECORD>
</NODE>

</HIERARCHY>

Working with Source Qualifiers

Source qualifiers determine how the PowerCenter Integration Service reads data from the source. You
connect a source qualifier to a source instance in a mapping. The type of source qualifier you use depends
on the source definition:

• Relational source definitions use a Source Qualifier transformation.

• Nonrelational source definitions use an Application Multi-Group Source Qualifier transformation.

Transformation Datatypes
The transformation datatypes in Source Qualifier and Application Multi-Group Source Qualifier
transformations are internal datatypes based on ANSI SQL-92 generic datatypes, which PowerCenter uses to
move data across platforms. When the PowerCenter Integration Service reads data from a source, it converts
the data from the PowerExchange datatype to the transformation datatype. When you run a session, the
PowerCenter Integration Service performs transformations based on the transformation datatypes. When
writing data to a target, the PowerCenter Integration Service converts the data based on the datatypes in the
target definition.

The transformation datatype for all ports in the Application Multi-Group Source Qualifier transformation are
predefined. You cannot change the datatype for any of the fields in the Application Multi-Group Source
Qualifier transformations.

Creating Source Qualifiers

By default, the PowerCenter Designer creates an Application Multi-Group Source Qualifier transformation
when you add a nonrelational source definition to a mapping. Similarly, it creates a Source Qualifier
transformation when you add a relational source definition to a mapping.

56 Chapter 3: Working with Mappings

 If you configure the PowerCenter Designer to manually create a source qualifier when you add a source
definition to a mapping, you must manually connect the source qualifier to the source definition.

Editing Source Qualifier Transformations

After you create a Source Qualifier transformation, you can edit the transformation. You cannot edit an
Application Multi-Group Source Qualifier transformation.

Lookup Transformations
This section discusses considerations for lookup transformations with PWXPC for PowerExchange sources
and targets.

Use a Lookup transformation in a mapping to look up data, through PowerExchange, in a nonrelational file or
a relational table, view, or synonym. You can import a lookup definition for any flat file, nonrelational file, or
relational table to which both the PowerCenter Client and PowerCenter Integration Service can connect.

Usage Considerations for PowerExchange Sources and Targets

Be aware of the following usage considerations for lookup transformations with PowerExchange sources and
targets:

• The Lookup transformation import process uses ODBC for nonrelational files and relational tables.
To use PWXPC to import definitions for nonrelational files or relational tables, first import the definitions in
the Import from PowerExchange dialog box in either the Source Analyzer or Target Designer before
configuring the lookup in the mapping.
• You can use PWXPC connections for lookup of both nonrelational files and relational tables:
- For relational tables, select the appropriate PWXPC relational connection for the database type, such as
PWX DB2zOS, PWX DB2i5OS, PWX DB2LUW, PWX Microsoft SQL Server, or PWX Oracle.
- For nonrelational files, select the PWXPC relational connection for NRDB lookups, PWX NRDB Lookup.
• For NRDB lookups, the Lookup transformation property Lookup Policy on Multiple Match has no effect.
The Lookup transformation always returns the first row that matches the lookup condition.
• To use lookup transformations with a resume recovery strategy, select the “Lookup source is static”
transformation attribute to avoid failures during session execution.
• To use lookup transformations with IMS databases, careful consideration needs to be given to the fields
used to perform the search of the IMS database. Concatenated keys (CCK) fields achieve the best
performance with the least impact on the IMS database.
• To use lookup transformations for targets being updated with CDC data in the same mapping, use special
custom properties to ensure change stream data is accessible across pipelines.

Related Topics:
• “NRDB Lookup Relational Connections” on page 90
• “Configuring Lookups for IMS” on page 58
• “Configuring Lookups for CDC Data” on page 59
• “Configuring PowerExchange Connections” on page 64

Lookup Transformations 57
 Limitations on Nonrelational Uncached Lookups
Nonrelational data sources that use the following PowerExchange access methods support cached lookups
but not uncached lookups:

• DB2UNLD
• DB2390IMG
• ESDS
• IMSUNLD
• RRDS
• TAPE
• USER
Nonrelational data sources that use the following PowerExchange access methods support both cached and
uncached lookups:

• ADABAS
• DATACOM
• DB2
• DL/1 BATCH (IMS)
• IDMS
• IMS ODBA
• KSDS
• SEQ

Configuring Lookups for IMS

You can use keys to retrieve data efficiently from IMS databases. PowerExchange enables you to build
Concatenated Key (CCK) fields in the data map. Concatenated Key (CCK) fields enables PowerExchange to
construct a fully-qualified Segment Search Argument (SSA) thereby improving IMS database search
efficiency.

Fields specified in the Lookup condition transformation attribute are used by PowerExchange to create the
Segment Search Argument (SSA). In order for a field to be used in the SSA, it must be marked as a key in
the IMS source or target definition in Designer.

To provide search keys for IMS database lookups, use the following types of fields in the Lookup condition
transformation attribute:

• Concatenated Key (CCK) fields. Specify these fields as keys in the IMS source or target definition and
use them in the Lookup condition attribute. The use of CCK fields results in the most efficient search of
the IMS database.
• Key fields. Specify these fields as key in the IMS source or target definition and use them in the Lookup
condition attribute. You can specify either the CCK field or the key field for the desired segment as both
will exist in the IMS source or target definition. If the segment is not the root, a combination of both CCK
fields and key fields will likely be needed in the Lookup condition.
• Search fields defined in the IMS Database Definition (DBD). Specify these as keys in the IMS source
or target definition and use them in the Lookup condition attribute. If the segment does not have a key,
IMS can still scan the IMS segments by using an IMS search field. This type of search call is not as
efficient as a keyed search with CCK fields or key fields. Assuming the root segment is keyed, include its

58 Chapter 3: Working with Mappings

 CCK field with the search fields to limit the amount of data IMS scans and therefore the impact on the
database.
• Non-key or non-search fields. The least efficient search method is to mark non-key fields or non-search
fields as keys in the IMS source or target definition and in the Lookup condition attribute. This causes a
scan of the IMS database to be done in order to find a match. This can adversely affect your IMS
operational system and therefore should be avoided.
Tip: You can limit the amount of the database that will be scanned by specifying as many CCK and key
fields as possible. If you use Search fields, include as many CCK fields as possible and, at minimum, the
root CCK field. Only use non-key or non-search fields as a last resort.
For more information about creating CCK fields in IMS data maps, see the PowerExchange Navigator User
Guide.

Configuring Lookups for CDC Data

Lookup transformations can be used in mappings extracting data from CDC sources including tables to which
the CDC data is being applied.

If you use dynamic lookups for CDC data and sharing the lookup cache with other static lookups in the same
mapping, specify both of the following custom properties in the session Custom Properties attribute field on
the Config Object tab:

• MergeCDCReaders=Yes
• SingleThreadExecutionModel=Yes
These custom properties remove any partition points from the PWXPC CDC Reader through the
transformations to the Writer. As a result, the order of the changes read from the change stream is
maintained until the changes reach the Writer. This then ensures that any CDC data placed into a dynamic
lookup cache is accessible to lookups sharing that cache in other pipelines.

Warning: The use of these custom properties will impact session throughput as it will single-thread all source
data from the Reader through to the Writer. As a result, these custom properties should only be specified
when there is a need to share CDC data stored in a dynamic cache across pipeline.

Stored Procedure Transformations

You can use Stored Procedure transformations that specify DB2 for z/OS or DB2 for i5/OS stored procedures
in a PowerCenter mapping. Use Stored Procedure transformations for read or write operations in bulk data
movement and change data capture (CDC) sessions.

Usage Considerations for Stored Procedure Transformations

The following usage considerations apply to Stored Procedure transformations for DB2 for z/OS and DB2 for
i5/OS:

• You must create and configure an ODBC data source before you import the DB2 for z/OS or DB2 for
i5/OS stored procedure.
• PowerCenter imports and calls stored procedures with unqualified procedure names that do not include a
schema name. DB2 executes unqualified stored procedures for PowerExchange by using the CURRENT
PATH register. DB2 searches the CURRENT PATH list of schema names from left to right until it finds a
schema name for which a stored procedure definition exists.

Stored Procedure Transformations 59

 However, PowerExchange, by default, executes unqualified stored procedures by using the User Name
value in the PWXPC connection as the schema name. For example, if the user name in the
PowerExchange connection is USER1, PowerExchange calls unqualified stored procedures as
USER1.procedure_name. Alternatively, for unqualified DB2 for i5/OS stored procedure names that are
used in bulk data movement, you can use the Library List attribute in the PowerCenter DB2i5OS
relational connection to qualify the stored procedure names.
Note: To explicitly set the schema name for a Stored Procedure transformation, edit the Stored
Procedure Name property of the Stored Procedure transformation.
• After you configure the Stored Procedure transformation in a mapping, use the PowerCenter Integration
Service to run the session.

Implementing a Stored Procedure Transformation in a Mapping

Implement a Stored Procedure transformation for a DB2 for z/OS or DB2 for i5/OS stored procedure in a
PowerCenter mapping.

1. Create and compile the DB2 stored procedure on the z/OS or i5/OS system.
2. Verify that a PowerExchange Listener is running on the z/OS or i5/OS system.
3. On the Windows system where the Mapping Designer runs, create and configure an ODBC data source
for the database that contains the stored procedure.
For more information, see the PowerExchange Reference Manual.
4. In the Mapping Designer, click Transformation > Import Stored Procedure.
5. In the ODBC data source box, enter or browse to the ODBC data source that you configured on the
Windows system.
6. In the Mapping Designer, add the Stored Procedure transformation to the mapping.
Note: When you import the stored procedure, the Mapping Designer does not qualify the stored
procedure name with the schema name. If you leave the name as is, PowerExchange executes any
unqualified stored procedures by using the User Name value in the PowerExchange connection as the
schema name. For unqualified DB2 for i5/OS stored procedure names that are used in PowerExchange
bulk data movement, you can use the Library List attribute of the DB2i5OS relational connection to
qualify the stored procedure name.
7. If the schema name that qualifies the stored procedure name is different from the user name under
which the PowerExchange Listener runs, and if you did not use the Library List attribute to qualify
DB2i5OS stored procedure names, edit the Stored Procedure transformation properties to specify the
schema name. In the Edit Transformations dialog box, click the Properties tab and add the schema
name before the stored procedure name in the Stored Procedure Name box.
For example, if the stored procedure name is ZEMPS1 and the user name under which the
PowerExchange Listener runs is USER2, enter USER2.ZEMPS in the Stored Procedure Name box.
For more information about Stored Procedure transformations, see the PowerCenter Transformation Guide.

60 Chapter 3: Working with Mappings

CHAPTER 4

Connections
This chapter includes the following topics:

• Connections Overview, 61
• Configuring PowerExchange Connections, 64
• DB2 Relational Connections, 65
• DB2 CDC Application Connections, 72
• NRDB Batch Application Connections, 79
• NRDB CDC Application Connections, 83
• NRDB Lookup Relational Connections, 90
• Microsoft SQL Server Relational Connections, 93
• Microsoft SQL Server CDC Application Connections, 96
• Oracle Relational Connections, 102
• Oracle CDC Application Connections, 105
• Configuring Connection Attributes, 111
• Connection Attributes for Multiple-Record Writes, 129
• Asynchronous Network Communication, 130

Connections Overview
Before the PowerCenter Integration Service can access a source or target in a session, you must configure
connections in the Workflow Manager. When you create or modify a session that reads from or writes to a
database or file, you can select only configured source and target databases. Connections are saved in the
repository.

For PowerExchange Client for PowerCenter, you configure relational or application connections, depending
on the source or the target type.

61
 Connection Types for Extracting Source Data
The following table describes the type of connections you can create to extract data from PowerExchange,
based on the data source and the extraction mode:

Source Type and Extraction Mode Connection Connection Type

Adabas bulk Application PWX NRDB Batch

Adabas batch from condense files Application PWX NRDB CDC Change

Adabas continuous from PowerExchange Logger for LUW log Application PWX NRDB CDC Real Time
files

Adabas real time Application PWX NRDB CDC Real Time

Datacom bulk Application PWX NRDB Batch

Datacom batch from condense files Application PWX NRDB CDC Change

Datacom continuous PowerExchange Logger for LUW log files Application PWX NRDB CDC Real Time

Datacom real time Application PWX NRDB CDC Real Time

DB2 for i5/OS bulk Relational PWX DB2i5OS

DB2 for i5/OS batch from condense files Application PWX DB2i5OS CDC Change

DB2 for i5/OS continuous from PowerExchange Logger for Application PWX DB2i5OS CDC Real Time
LUW log files

DB2 for i5/OS real time Application PWX DB2i5OS CDC Real Time

DB2 for Linux, UNIX, and Windows bulk Relational PWX DB2LUW

DB2 for Linux, UNIX, and Windows batch from Application PWX DB2LUW CDC Change
PowerExchange Logger for LUW log files

DB2 for Linux, UNIX, and Windows continuous from Application PWX DB2LUW CDC Real Time
PowerExchange Logger for LUW log files

DB2 for Linux, UNIX, and Windows real time Application PWX DB2LUW CDC Real Time

DB2 for z/OS bulk Relational PWX DB2zOS

DB2 for z/OS batch from condense files Application PWX DB2zOS CDC Change

DB2 for z/OS continuous from PowerExchange Logger for Application PWX DB2zOS CDC Real Time
LUW log files

DB2 for z/OS real time Application PWX DB2zOS CDC Real Time

IDMS bulk Application PWX NRDB Batch

IDMS batch from condense files Application PWX NRDB CDC Change

62 Chapter 4: Connections
Source Type and Extraction Mode Connection Connection Type

IDMS continuous from PowerExchange Logger for LUW log Application PWX NRDB CDC Real Time
files

IDMS real time Application PWX NRDB CDC Real Time

IMS bulk Application PWX NRDB Batch

IMS batch from condense files Application PWX NRDB CDC Change

IMS continuous from PowerExchange Logger for LUW log files Application PWX NRDB CDC Real Time

IMS real time Application PWX NRDB CDC Real Time

Microsoft SQL Server bulk Relational PWX MSSQLServer

Microsoft SQL Server batch from PowerExchange Logger for Application PWX MSSQL CDC Change
LUW log files

Microsoft SQL Server continuous from PowerExchange Logger Application PWX MSSQL CDC Real Time
for LUW log files

Microsoft SQL Server real time Application PWX MSSQL CDC Real Time

Oracle bulk Relational PWX Oracle

Oracle batch from PowerExchange Logger for LUW log files Application PWX Oracle CDC Change

Oracle continuous from PowerExchange Logger for LUW log Application PWX Oracle CDC Real Time
files

Oracle real time Application PWX Oracle CDC Real Time

Sequential and flat files bulk Application PWX NRDB Batch

Sequential and flat files batch from condense files n/a n/a

Sequential and flat files real time n/a n/a

VSAM bulk Application PWX NRDB Batch

VSAM batch from condense files Application PWX NRDB CDC Change

VSAM continuous from PowerExchange Logger for LUW log Application PWX NRDB CDC Real Time
files

VSAM real time Application PWX NRDB CDC Real Time

Connections Overview 63
 Connection Types for Loading Target Data
The following table describes the types of relational and application connections that you can define to load
data to PowerExchange targets:

Target Type Connection Connection Type

Adabas bulk Application PWX NRDB Batch

DB2 for i5/OS bulk Relational PWX DB2i5OS

DB2 for Linux, UNIX, and Windows bulk Relational PWX DB2LUW

DB2 for z/OS bulk Relational PWX DB2zOS

IMS bulk Application PWX NRDB Batch

Microsoft SQL Server bulk Relational PWX MSSQLServer

Oracle bulk Relational PWX Oracle

Sequential and flat files bulk Application PWX NRDB Batch

VSAM bulk Application PWX NRDB Batch

Configuring PowerExchange Connections

Configure PWXPC connections to data sources and targets. Use the connections in sessions that extract
bulk data or change data from a relational or nonrelational source or that load bulk data to a relational or
nonrelational target.

1. In the Workflow Manager, connect to a PowerCenter repository.

2. To configure a PWXPC relational connection, click Connections > Relational.
The Relational Connection Browser dialog box appears.
3. To configure a PWXPC application connection, click Connections > Application.
The Application Connection Browser dialog box appears.
4. In the Select Type field, select the specific type of connection.
5. Click New.
The Connection Object Definition dialog box appears.
6. Enter values for the connection attributes.
For more information, see the topic for the specific connection type.
7. Click OK.
The new connection appears in the Application or Relational Object Browser.

Related Topics:
• “DB2 CDC Application Connections” on page 72
• “DB2 Relational Connections” on page 65

64 Chapter 4: Connections
 • “Microsoft SQL Server CDC Application Connections” on page 96
• “Microsoft SQL Server Relational Connections” on page 93
• “NRDB Batch Application Connections” on page 79
• “NRDB CDC Application Connections” on page 83
• “NRDB Lookup Relational Connections” on page 90
• “Oracle CDC Application Connections” on page 105
• “Oracle Relational Connections” on page 102

DB2 Relational Connections

Configure relational connections to extract data from and load data to DB2 for z/OS, DB2 for i5/OS, and DB2
for Linux, UNIX, and Windows sources and targets in bulk data movement sessions. Also use relational
connections to apply change data to DB2 relational targets in CDC sessions.

The following table describes the connection attributes for the PWX DB2zOS, PWX DB2i5OS, and PWX
DB2LUW relational connection types:

Connection Required DB2zOS, Description

Attribute DB2i5OS,
DB2LUW
Types

Name Yes All The name of the relational connection.

Code Page Yes All The code page that the PowerCenter Integration Service uses to
extract data from the source database.
Note: In Unicode mode, PWXPC sets the code page with this value
overriding any code page specification in the PowerExchange
dbmover configuration file.

Location Yes All The location of the source or target database as specified as
specified in a NODE statement in the PowerExchange dbmover
configuration file.

Database Yes All The DB2 subsystem or database name.

Name

User Name Yes All A user name that can be used to access the database.
For databases on supported Linux, UNIX, or Windows systems, if
you have enabled PowerExchange LDAP user authentication and
disabled relational pass-through authentication, the user name is
the enterprise user name. For more information, see the
PowerExchange Reference Manual.

DB2 Relational Connections 65

 Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

Password Yes All A password for the specified user. Passwords that are less than
nine characters in length are not validated.
To avoid errors that might arise due to code page differences,
Informatica recommends that the password includes only those
characters that are allowed for passphrases.
For DB2i5OS and DB2zOS relational connections, you can enter a
valid PowerExchange passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length when you use
a PWXPC connection. A z/OS passphrase can be from 9 to 128
characters in length. A passphrase can contain the following
characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | <
> ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double
quotation marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do
not affect the allowable characters in PowerExchange passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100
characters in length. PowerExchange truncates passphrases longer
than 100 characters when passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs
with a security setting of SECURITY=(1,N) or higher in the
DBMOVER member. For more information, see "SECURITY
Statement" in the PowerExchange Reference Manual.

Isolation Yes DB2i5OS The commit scope of the transaction.

Level Valid values:
- None
- CS. Cursor stability.
- RR. Repeatable Read.
- CHG. Change.
- ALL
Default is CS.

66 Chapter 4: Connections
Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

Database file No DB2i5OS Override for the i5/OS database file. The format is:
overrides
from_file/to_library/to_file/to_member
Where:
- from_file is the file to be overridden.
- to_library is the new library to use.
- to_file is the file in the new library to use.
- to_member is optional and is the member in the new library and file
to use. If no value is specified, *FIRST is used.
You can enter up to eight unique file overrides on a single
connection. A single override applies to a single source or target.
When you specify multiple file overrides, include a space between
each file override and enclose the entire string of file overrides in
double quotation marks.
Note: If both Library List and Database file overrides are
specified and a table exists in both, Database file overrides takes
precedence.

Library List No DB2i5OS A list of the libraries that PowerExchange searches to qualify a
table name for Select, Insert, Delete, or Update statements or to
qualify a stored procedure name when the procedure is called for
execution. PowerExchange searches the list if a table name or
stored procedure name is not qualified with a schema name.
If you specify multiple libraries, separate the library names with
spaces and enclose the list in double quotation marks, for example,
"TGTLIB1 TGTLIB2 TGTLIB3".
Note: If both Library List and Database file overrides are
specified and a table exists in both, Database file overrides takes
precedence.

Environment No All SQL commands that run in the database environment.

SQL

Compression No All Select this option to compress source data during the PowerCenter
session.

Encryption Yes All The encryption type. For information about supported encryption
Type types, see “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No All The encryption level. For information about supported encryption
Level levels, see “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size Yes All The pacing size. For information about setting the pacing size, see
“Configuring Pacing” on page 113 .
Default is 0.

Interpret as No All Controls whether the pacing size is in number of rows.

Rows

DB2 Relational Connections 67

 Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

Bulk Load No DB2zOS Controls whether PowerExchange loads data to DB2 for z/OS
targets with the DB2 LOAD utility. If you select this option, you can
configure the remaining connection attributes that apply to the DB2
LOAD utility. Otherwise, PowerExchange ignores these attributes.

Filename No DB2zOS The data set prefix that PowerExchange uses to create the
temporary files that the DB2 LOAD utility uses to load data to a
DB2 table.

Space Yes DB2zOS Controls whether to allocate z/OS space in tracks or cylinders.
Valid values:
- TRACK
- CYLINDER
Default is TRACK.

Primary Yes DB2zOS The primary space on z/OS to use for LOAD operations. Default is
Space 0.

Secondary Yes DB2zOS The secondary space on z/OS to use for LOAD operations. Default
Space is 0.

Delete Yes DB2zOS Determines how PowerExchange handles the temporary files it
Temporary creates for the DB2 LOAD utility to load data into a DB2 table.
Files Valid values:
- NO. Do not delete the temporary files.
- BEFORE. Delete the temporary files before running the utility.
- AFTER SUCCESS ONLY. Delete the temporary files if the utility
ends successfully with return code 0.
- AFTER. Delete the temporary files after the utility runs.
Default is NO.

JCL Template Yes DB2zOS The name of the JCL template for the DB2 LOAD utility on the
PowerExchange target system.
Default is DB2LDJCL.

CTL Template Yes DB2zOS The name of the control file template for the DB2 LOAD utility on
the PowerExchange on the target system.
Default is DB2LDCTL.

Load Options Yes DB2zOS Indicates how the data that PowerExchange provides to the DB2
LOAD utility is loaded into a DB2 table.
Valid values:
- RESUME. Generate a LOAD RESUME statement.
- REPLACE. Generate a LOAD REPLACE statement.
Default is RESUME.

68 Chapter 4: Connections
Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

Mode Type Yes DB2zOS Indicates how PowerExchange runs the DB2 LOAD utility to load
data into a DB2 table.
Valid values:
- TASK. Run the LOAD utility as a subtask under the PowerExchange
Listener.
- JOB. Submit a separate job to run the DB2 LOAD utility.
- NOSUBMIT. Create the files and JCL to run the DB2 LOAD utility
(unless Mode Time is set to DATAONLY) but do not submit the load
job. You must submit the job manually.
Default is TASK.

Mode Time Yes DB2zOS Determines how PowerExchange handles the execution of the DB2
LOAD utility.
Valid values:
- WAIT. Wait for the job to end before returning control to
PowerCenter. This option can be specified only when the Mode Type
is JOB or TASK.
- NO WAIT. Return to PowerCenter without waiting for the job to end.
This option can be specified only when Mode Type is set to JOB or
NOSUBMIT.
- TIMED. Wait the number of seconds that is specified in the Time
attribute before returning control to PowerCenter. The TIMED option
can be specified only when Mode Type is set to JOB.
- DATAONLY. Create the data file only. Do not create the files and
JCL for running the DB2 LOAD utility. Usually, this option is used
when Mode Type is set to NOSUBMIT.
Default is WAIT.
Note: If you enter WAIT, PowerExchange uses 99,999 seconds as
the network operation timeout value instead of the value specified
by the TCPIP_OP_TIMEOUT parameter of the PWX Override
connection attribute. If you enter TIMED, PowerExchange adds 5
minutes to the network operation timeout value specified by the
connection attribute.

Time Yes DB2zOS The wait time, in seconds, that is in effect when you set Mode Type
to JOB and Mode Time to TIMED.
Valid values are 1 to 99998.
Default is 0.

Convert No All Controls whether character fields are converted to string fields so
character that embedded nulls in data are processed as spaces.
data to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on
page 114 .

Write Mode No All The write mode. For information about Write Mode options, see
“Configuring Write Mode” on page 115 .
Default is Confirm Write On.

DB2 Relational Connections 69

 Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

Reject File No All Overrides the default prefix of PWXR for the reject file.
PowerExchange creates the reject file on the target machine when
the Write Mode is Asynchronous with Fault Tolerance.
Note: Enter PWXDISABLE to prevent creation of the reject files.

Correlation Id No DB2zOS A value to use as the DB2 Correlation ID for DB2 requests.
This value overrides the value in the SESSID statement in the
PowerExchange DBMOVER configuration file.

Offload No DB2zOS Indicates whether to use offload processing to move

Processing PowerExchange bulk data processing from the source system to
the PowerCenter Integration Service machine.
Valid values:
- No. Do not use offload processing.
- Yes. Use offload processing.
- Auto. PowerExchange determines whether to use offload
processing.
Default is No.

Worker No DB2zOS When offload processing is enabled, specifies the number of

Threads threads that PowerExchange uses on the PowerCenter Integration
Service machine to process bulk data.
For optimal performance, this value should not exceed the number
of installed or available processors on the PowerCenter Integration
Service machine.
Valid values are 1 through 64.
Default is 0, which disables multithreading.

Array Size No DB2zOS, The DB2 fetch array size, in number of rows, for DB2 bulk data
DB2i5OS movement operations that use the DB2 access method. The array
size pertains to the DB2 multiple-row FETCH statements that
PowerExchange uses to read data from DB2 source tables.
Valid values are from 1 through 100000. Default is 25.
Tip: For DB2 for i5/OS bulk data movement operations, Informatica
recommends that you enter an array size no greater than 100.
Note: PowerExchange dynamically lowers the array size when all
the following conditions are true:
- The table contains LOB columns.
- The Array Size value is greater than 1.
- Row size * Array Size is greater than 16000000 bytes.

70 Chapter 4: Connections
Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

PWX Override No All PowerExchange connection overrides, separated by a semicolon.

You can enter one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the
application buffer for the connection. Overrides the APPBUFSIZE
statement in the DBMOVER configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the
characters in the connection string. Enter this override if
PowerExchange issues message PWX-07610.
- CONNSHARE={N|Y}. By default, all DB2 lookups in a workflow use
the same connection, and the PowerExchange Listener performs
them in a single task.
To preserve the behavior in effect for DB2 lookups in
PowerExchange releases earlier than 9.6.0, specify
CONNSHARE=N.
Caution: If the DB2 connection is used as a target in a CDC
workflow, do not change the default behavior. Otherwise, internal
PowerCenter state tables that require connection sharing might not
be updated correctly.
Note: DB2 lookups that share a connection do not use offload
processing, partitioning, or threading.
- DB2TRUNCASDEL={OFF|ON}. For DB2zOS target connections, by
default, if the PowerCenter Integration Service requests a
TRUNCATE statement, PowerExchange issues a TRUNCATE
statement.
If DB2TRUNCASDEL=ON, when the PowerCenter Integration
Service requests a TRUNCATE statement, PowerExchange issues a
DELETE statement.
Note: The DB2TRUNCASDEL override applies only to DB2zOS
target connections. For DB2i5OS target connections,
PowerExchange always issues a DELETE statement. For DB2LUW
target connections, PowerExchange always issues a TRUNCATE
statement.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves
hexadecimal '0' values, called low values, in source character fields
when passing these values to a PowerCenter session. Enter Y to
preserve low values. When the session runs, PWXPC can write
these values to a VSAM target on z/OS or to a sequential file target
on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES
statement in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements
that PWXPC uses to select data for session processing, PWXPC
adds double-quotation marks (") around every table name and field
name that includes spaces or special characters. These special
characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only
around the "table_name" value.
If for any reason, you cannot use quoted table names or field names
in your environment, set PWXNOQUOTES to Y to override the
default behavior.

DB2 Relational Connections 71

 Connection Required DB2zOS, Description
Attribute DB2i5OS,
DB2LUW
Types

- RETLOGINFOMSG={N|Y}. By default, PWXPC writes

PowerExchange error and warning messages but not informational
messages to the session log. If you specify RETLOGINFOMSG=Y,
PWXPC writes informational messages as well as error and warning
messages to the session log.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network
operation timeout interval in seconds. For more information about
this timeout override, see “Asynchronous Network
Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout
interval in seconds. For more information about this timeout override,
see “Asynchronous Network Communication” on page 130.

Connection No All The number of seconds the PowerCenter Integration Service

Retry Period attempts to reconnect to the PowerExchange Listener after the
initial connection attempt fails. If the Integration Service cannot
connect to the PowerExchange Listener within the retry period, the
session fails.
Default value is 0, which disables connection retries based on this
PWXPC connection attribute.
Note: PowerExchange Listener connections to source databases
do not have connection resilience.

Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Configuring Write Mode” on page 115

DB2 CDC Application Connections

Configure CDC application connections to extract change data from DB2 for z/OS, DB2 for i5/OS, or DB2 for
Linux, UNIX, and Windows sources in real-time extraction mode. Also configure CDC application connections
to extract change data from PowerExchange Condense files in batch extraction mode or from
PowerExchange Logger for Linux, UNIX, and Windows log files in batch or continuous extraction mode.

Depending on the DB2 source type, you can configure the following types of application connections:

• PWX DB2zOS CDC Change

• PWX DB2zOS CDC Real Time
• PWX DB2i5OS CDC Change
• PWX DB2i5OS CDC Real Time
• PWX DB2LUW CDC Change
• PWX DB2LUW CDC Real Time

72 Chapter 4: Connections
The following table describes the connection attributes:

Connection Required? Change Description

Attribute or Real
Time

Name Yes Both The name of the application connection.

Code Page Yes Both The code page that the PowerCenter Integration Service uses to
extract data from the source database.
In Unicode mode, you must select UTF-8 for DB2LUW CDC
connections.
Note: In Unicode mode, PWXPC sets the code page with this value
overriding any code page specification in the PowerExchange
configuration file.

Location Yes Both The location of the source or target database as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file.

User Name Yes Both A user name that can be used to access the DB2 database.
For a database on a supported Linux, UNIX, or Windows system, if
you have enabled PowerExchange LDAP user authentication, the
user name is the enterprise user name. For more information, see
the PowerExchange Reference Manual.

Password Yes Both A password for the specified user. Passwords that are less than
nine characters in length are not validated.
To avoid errors that might arise due to code page differences,
Informatica recommends that the password includes only those
characters that are allowed for passphrases.
For PWX DB2i5OS and DB2zOS CDC, you can enter a valid
PowerExchange passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length. A z/OS
passphrase can be from 9 to 128 characters in length when you
use a PWXPC connection. A passphrase can contain the following
characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | <
> ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double
quotation marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do
not affect the allowable characters in PowerExchange
passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100
characters in length. PowerExchange truncates passphrases
longer than 100 characters when passing them to RACF for
validation.
To use passphrases, ensure that the PowerExchange Listener runs
with a security setting of SECURITY=(1,N) or higher in the
DBMOVER member. For more information, see "SECURITY
Statement" in the PowerExchange Reference Manual.

DB2 CDC Application Connections 73

 Connection Required? Change Description
Attribute or Real
Time

Compression No Both Select this option to compress source data during the PowerCenter
session.

Encryption Yes Both The encryption type. For information about encryption types, see
Type “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No Both The encryption level. For information about encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size Yes Both The pacing size. For information about setting the pacing size, see
“Configuring Pacing” on page 113 .
Default is 0.

Interpret as No Both Controls whether the pacing size is in number of rows.

Rows

Image Type No Both Indicates whether PWXPC passes captured Update operations to
CDC sessions as Updates, or as Deletes followed by Inserts.
Default is BA, which processes Updates as Deletes followed by
Inserts.
For more information, see “Configuring Image Type” on page 117 .

Application No Both The application name.

Name

RestartToken Yes Both The restart token folder.

File Folder Default is $PMRootDir/Restart.
For more information, see “Configuring CDC Restart Attributes” on
page 121 .

RestartToken No Both The restart token file name.

File Name

Number of No Both The maximum number of backup copies of the restart token file to
Runs to Keep keep.
RestartToken Default is 0.
File

Recovery No Both The file cache folder to enable recovery for the session.
Cache Folder Default is $PMRootDir/Cache.

74 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

UOW Count No Both The number of units of work (UOWs) that PWXPC processes
before it sends a commit request to PowerCenter to commit data to
all targets in the CDC session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before
sending a commit request.
Default is 1.
For more information, see “Configuring UOW Count” on page 121 .

Reader Time No Real The number of seconds that the PowerCenter Integration Service
Limit Time reads data from the source before stopping.
If you enter 0, the Reader Time Limit attribute does not limit the
reader time. This attribute is intended for testing purposes only.
Tip: Use Idle Time instead of Reader Time Limit.
Default is 0.

Idle Time No Real The number of seconds the PowerExchange Listener remains idle
Time after reaching the end of the change log, as indicated by message
PWX-09967, before returning an end-of-file (EOF).
Valid values:
-1. The EOF is not returned. The session runs continuously.
0. The EOF is returned at the end of log. The session ends
successfully.
n . The number of seconds after which the EOF is returned.
Default is -1.
For more information, see “Configuring Idle Time” on page 119 .

Real-Time No Real The number of milliseconds that must elapse before PWXPC sends
Flush Latency Time a commit request to PowerCenter. When this period elapses,
in milli- PWXPC continues to read the changes in the current UOW until it
seconds reaches the end of the UOW. Then PWXPC sends a commit
request to PowerCenter so that the data can be committed to the
targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse
before PWXPC sends a commit request to PowerCenter. If you
specify a value from 0 through 2000, PWXPC uses 2000
milliseconds (2 seconds).
Default is 0.
For more information, see “Configuring PWX Latency and Real-Time
Flush Latency” on page 122 .

PWX Latency No Real The maximum number of seconds that PowerExchange extraction
in seconds Time processing on the source platform waits for more change data
before it returns control to PWXPC on the PowerCenter Integration
Service machine.
Default is 2.

DB2 CDC Application Connections 75

 Connection Required? Change Description
Attribute or Real
Time

Maximum No Both The maximum number of change records that PWXPC processes
Rows Per before it sends a commit request to PowerCenter to commit data to
commit all targets in the CDC session. PWXPC does not wait for a UOW
boundary to commit the change data.
Default is 0, which disables this attribute.
For more information, see “Configuring Maximum Rows per
Commit” on page 124 .

Minimum Rows No Real The minimum number of change records that the PowerExchange
Per commit Time Listener must read from the change stream before it passes a
commit record to PWXPC. Before reaching this minimum, the
PowerExchange Listener passes only change records to PWXPC
and discards any commit records.
Default is 0, which disables this attribute.
For more information, see “Configuring Minimum Rows per
Commit” on page 125 .

Journal Name No Both Overrides the library and journal name in the PowerExchange
Note: For PWX CAPI_CONNECTION.
DB2i5OS CDC Specify complete library and journal names in the format:
Real Time
application library/journal
connections
only.

Library/File No Both Overrides the library and file name in the extraction maps of all
Override sources that use the connection.
Note: For PWX Specify a single library and file name in the format:
DB2i5OS CDC
Real Time library/file
application If specifying asterisk for the library name, use the session
connections properties Library/File Override attribute.
only. Note: This attribute is overridden by the Library/File Override value
on the session properties.

Convert No Both Controls whether to convert character fields to string fields so that
character data embedded nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on
page 114 .

Event Table No Real The PowerExchange extraction map name that is used for event
Time processing.
For more information, see “Configuring Event Table Processing” on
page 118 .

76 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

Offload No Real Controls whether PowerExchange uses CDC offload processing.

Processing Time When offload processing is enabled, PowerExchange transfers
column-level processing of the change data from the source
system to the PowerCenter Integration Service machine.
Valid values:
- No. Disables offload processing.
- Yes. Enables offload processing.
- Auto. PowerExchange determines whether to use offload
processing.
Default is No.
For more information about offload processing, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Worker No Real If you use CDC offload processing, the number of threads that
Threads Time PowerExchange uses on the PowerCenter Integration Service
machine to process change data.
For optimal performance, this value should not exceed the number
of installed or available processors on the Integration Service
machine.
Valid values are from 1 through 64.
Default is 0.

Array Size No Real If the Worker Threads value is greater than 0, specifies the size of
Time the storage array, in number of records, for the threads.
Valid values are from 1 through 100000.
Default is 25.

Map Location No Real If you use remote logging to a PowerExchange Logger for Linux,
Time UNIX, and Windows, the location of the source extraction maps.
This value must be the node name of the location, as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file on the Integration Service machine.
For more information about remote logging, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Map Location No Real If you specify a Map Location value, a valid user ID for the map
User Time location.

Map Location No Real If you specify a Map Location value, a valid password for the
Password Time specified map location user.

CAPI No Real Overrides the default CAPI connection name.

Connection Time For more information, see “Configuring the CAPI Connection Name
Name Override Override” on page 119 .

Retrieve PWX No Both Controls whether PWXPC writes PowerExchange informational

Log Entries messages related to a CDC session in the PowerCenter session
log.
Default is to write PowerExchange error and warning messages but
not informational messages to the session log.

DB2 CDC Application Connections 77

 Connection Required? Change Description
Attribute or Real
Time

PWX Override No Both PowerExchange connection overrides, separated by a semicolon.

You can enter one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the
application buffer for the connection. Overrides the APPBUFSIZE
statement in the DBMOVER configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the
characters in the connection string. Enter this override if
PowerExchange issues message PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves
hexadecimal '0' values, called low values, in source character fields
when passing these values to a PowerCenter session. Enter Y to
preserve low values. When the session runs, PWXPC can write
these values to a VSAM target on z/OS or to a sequential file target
on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES
statement in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements
that PWXPC uses to select data for session processing, PWXPC
adds double-quotation marks (") around every table name and field
name that includes spaces or special characters. These special
characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only
around the "table_name" value.
If for any reason, you cannot use quoted table names or field names
in your environment, set PWXNOQUOTES to Y to override the
default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes
PowerExchange error and warning messages but not informational
messages to the session log. If you specify RETLOGINFOMSG=Y,
PWXPC writes informational messages as well as error and warning
messages to the session log.
Note: The RETLOGINFOMSG connection override is equivalent to
the Retrieve PWX Log Entries connection attribute.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network
operation timeout interval in seconds. For more information about
this timeout override, see “Asynchronous Network
Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout
interval in seconds. For more information about this timeout
override, see “Asynchronous Network Communication” on page
130.

Connection No Both The number of seconds the PowerCenter Integration Service

Retry Period attempts to reconnect to the PowerExchange Listener after the
initial connection attempt fails. If the Integration Service cannot
connect to the PowerExchange Listener within the retry period, the
session fails.
Default value is 0, which disables connection retries based on this
PWXPC connection attribute.
Note: PowerExchange Listener connections to source databases
do not have connection resilience.

78 Chapter 4: Connections
 Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Retrieving PowerExchange Log Entries” on page 115
• “Configuring Image Type” on page 117
• “Configuring the CAPI Connection Name Override” on page 119
• “Configuring Idle Time” on page 119
• “Configuring CDC Restart Attributes” on page 121
• “Configuring UOW Count” on page 121
• “Configuring PWX Latency and Real-Time Flush Latency” on page 122
• “Configuring Maximum Rows per Commit” on page 124
• “Commit Processing for CDC Sessions” on page 161
• “Managing Session Log and Restart Token File History” on page 185
• “Configuring Event Table Processing” on page 118
• “Enabling Session Recovery for a CDC Session” on page 177

NRDB Batch Application Connections

Configure NRDB Batch application connections to extract data from and load data to nonrelational sources
and targets in bulk data movement sessions. Also use NRDB Batch application connections to apply change
data to nonrelational targets in CDC sessions.

The following table describes the connection attributes for the PWX NRDB Batch application connection type:

Connection Required Description

Attribute

Name Yes The name of the application connection.

Code Page Yes The code page for the PowerCenter Integration Service to use to extract data
from the source database.
Note: In Unicode mode, PWXPC sets the code page with this value overriding
any code page specification in the PowerExchange configuration file.

Location Yes The location of the source or target database as specified in a NODE statement in
the PowerExchange dbmover.cfg configuration file.

User Name Yes The database user name.

For databases on supported Linux, UNIX, or Windows systems, if you have
enabled PowerExchange LDAP user authentication, the user name is the
enterprise user name. For more information, see the PowerExchange Reference
Manual.

NRDB Batch Application Connections 79

 Connection Required Description
Attribute

Password Yes A password for the specified user or a valid PowerExchange passphrase.
Passwords that are less than nine characters in length are not validated.
To avoid errors that might arise due to code page differences, Informatica
recommends that the password includes only those characters that are allowed
for passphrases.
For nonrelational sources or targets on z/OS, and for sequential source or target
files on i5/OS, you can enter a passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length. A z/OS passphrase can be
from 9 to 128 characters in length with a PWXPC connection. A passphrase can
contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation marks
(“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not affect the
allowable characters in PowerExchange passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in length.
PowerExchange truncates passphrases longer than 100 characters when passing
them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member. For
more information, see "SECURITY Statement" in the PowerExchange Reference
Manual.
To use passphrases for IMS connections, ensure that the following additional
requirements are met:
- You must configure ODBA access to IMS as described in the PowerExchange
Navigator User Guide.
- You must use IMS data maps that specify IMS ODBA as the access method. Do not
use data maps that specify the DL/1 BATCH access method because this access
method requires the use of netport jobs, which do not support passphrases.
- The IMS database must be online in the IMS control region to use ODBA access to
IMS.

Compression No Controls whether to compress source data during the session.

Encryption No The encryption type. For information about encryption types, see “Configuring
Type Encryption and Compression” on page 112 .
Default is None.

Encryption No The encryption level. For information about encryption levels, see “Configuring
Level Encryption and Compression” on page 112 .
Default is 1.

Pacing Size No The pacing size. For information about setting the pacing size, see “Configuring
Pacing” on page 113 .
Default is 0.

80 Chapter 4: Connections
Connection Required Description
Attribute

Interpret as No Controls whether the pacing size is expressed in kilobytes or number of rows.
Rows

Offload No Controls whether to use offload processing. Offload processes transfers

Processing PowerExchange bulk data processing from the source system to the PowerCenter
Integration Service machine.
Valid values:
- No. Disables offload processing.
- Auto. PowerExchange determines whether to use offload processing.
- Filter Before. Offloads processing to the PowerCenter Integration Service machine
but continues to filter data on the source system.
If you select Filter Before for an IMS unload file data source, PowerExchange
changes the selection to Filter After.
- Filter After. Offloads the bulk data processing to the PowerCenter Integration
Service machine, including the filtering of data.
Default is No.

Worker No The number of threads that PowerExchange uses on the PowerCenter Integration
Threads Service machine to process bulk data.
For optimal performance, this value should not exceed the number of installed or
available processors on the Integration Service machine.
Valid values are from 1 through 64.
Default is 0, which disables multithreading. If you use reader or writer pipeline
partitioning, accept the default value of 0. You cannot use both multiple worker
threads and partitioning.

Array Size No The storage array size, in number of records, that is used for partitioned or
multithreaded sessions.
For partitioned sessions, this array size is shared across the partitions.
Valid values are from 1 through 100000. Default is 25.
To tune partitioned sessions, particularly when the Write Mode attribute specifies
Confirm Write On, increase the array size.

Write Mode No Controls whether data is sent to the PowerExchange Listener synchronously or
asynchronously.
For information about Write Mode options, see “Configuring Write Mode” on page
115 .
Default is Confirm Write On.

Convert No Controls whether character fields are converted to string fields so that embedded
character data nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on page 114 .

Retrieve PWX No Controls whether PWXPC writes PowerExchange informational messages related
Log Entries to the session in the PowerCenter session log.
Default is to write PowerExchange error and warning messages but not
informational messages to the session log.

NRDB Batch Application Connections 81

 Connection Required Description
Attribute

PWX Override No PowerExchange connection overrides, separated by a semicolon. You can enter
one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the application buffer for
the connection. Overrides the APPBUFSIZE statement in the DBMOVER
configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the characters in the
connection string. Enter this override if PowerExchange issues message
PWX-07610.
- CSQ_MEMSIZE=memory_size. The memory size, in megabytes, for the cache that
is used for multiple-record writes before it spills to disk. Default is 16.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves hexadecimal '0'
values, called low values, in source character fields when passing these values to a
PowerCenter session. Enter Y to preserve low values. When the session runs,
PWXPC can write these values to a VSAM target on z/OS or to a sequential file
target on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES statement in
the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements that PWXPC
uses to select data for session processing, PWXPC adds double-quotation marks (")
around every table name and field name that includes spaces or special characters.
These special characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only around the
"table_name" value.
If for any reason, you cannot use quoted table names or field names in your
environment, set PWXNOQUOTES to Y to override the default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes PowerExchange error and
warning messages but not informational messages to the session log. If you specify
RETLOGINFOMSG=Y, PWXPC writes informational messages as well as error and
warning messages to the session log.
Note: The RETLOGINFOMSG connection override is equivalent to the Retrieve
PWX Log Entries connection attribute.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network operation timeout
interval in seconds. For more information about this timeout override, see
“Asynchronous Network Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout interval in
seconds. For more information about this timeout override, see “Asynchronous
Network Communication” on page 130.
- WRT_ERROR_HANDLING=Y. If you use writer partitioning and set the Write Mode
attribute to Confirm Write On, specify WRT_ERROR_HANDLING=Y to use an
alternative method of error handling that is more efficient than the default method
when the input data contains many errors. You can also specify this statement in
the DBMOVER configuration file.

Connection No The number of seconds the PowerCenter Integration Service attempts to

Retry Period reconnect to the PowerExchange Listener after the initial connection attempt fails.
If the Integration Service cannot connect to the PowerExchange Listener within
the retry period, the session fails.
Default value is 0, which disables connection retries based on this PWXPC
connection attribute
Note: PowerExchange Listener connections to source databases do not have
connection resilience.

Default value is 0.

82 Chapter 4: Connections
 Related Topics:
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Configuring Write Mode” on page 115
• “Retrieving PowerExchange Log Entries” on page 115

NRDB CDC Application Connections

Configure NRDB CDC application connections to extract change data from nonrelational sources in real-time
extraction mode or from PowerExchange Condense condense files in batch extraction mode.

The following table describes the connection attributes for the PWX NRDB CDC Change and CDC Real Time
application connection types:

Connection Required? Change Description

Attribute or Real
Time

Name Yes Both A name for the application connection.

Code Page Yes Both The code page that the PowerCenter Integration Service uses to
extract data from the source database.
Note: In Unicode mode, PWXPC sets the code page with this value
overriding any code page specification in the PowerExchange
configuration file.

Location Yes Both The location of the source or target database, as specified in a
NODE statement in the DBMOVER configuration file.

User Name Yes Both A user name for connecting to the database.
For connections to supported Linux, UNIX, or Windows platforms, if
you have enabled PowerExchange LDAP user authentication, the
user name is the enterprise user name. For more information, see
the PowerExchange Reference Manual.

NRDB CDC Application Connections 83

 Connection Required? Change Description
Attribute or Real
Time

Password Yes Both A password for the specified user or a valid PowerExchange
passphrase.
Passwords that are less than nine characters in length are not
validated.
To avoid errors that might arise due to code page differences,
Informatica recommends that the password includes only those
characters that are allowed for passphrases.
A PowerExchange passphrase for z/OS access can be from 9 to
128 characters in length when you use a PWXPC connection. A
passphrase can contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | <
> ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double
quotation marks (“), or currency symbols.
The allowable characters in the IBM IRRPHREX exit do not affect
the allowable characters in PowerExchange passphrases.
Note: A valid RACF passphrase can be up to 100 characters in
length. PowerExchange truncates passphrases longer than 100
characters when passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs
with a security setting of SECURITY=(1,N) or higher in the
DBMOVER member. For more information, see "SECURITY
Statement" in the PowerExchange Reference Manual.
To use passphrases for IMS connections, ensure that the following
additional requirements are met:
- You must configure ODBA access to IMS as described in the
PowerExchange Navigator User Guide.
- You must use IMS data maps that specify IMS ODBA as the access
method. Do not use data maps that specify the DL/1 BATCH access
method because this access method requires the use of netport
jobs, which do not support passphrases.
- The IMS database must be online in the IMS control region to use
ODBA access to IMS.

Compression No Both Controls whether to compress source data during the PowerCenter
session.

Encryption Yes Both The encryption type. For information about supported encryption
Type types, see “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No Both The encryption level. For information about supported encryption
Level levels, see “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size Yes Both The pacing size.

84 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

Interpret as No Both Indicates whether the pacing size is in number of rows.

Rows

Image Type No Both Indicates how PWXPC passes captured Updates to CDC sessions:
as Updates, or as Deletes followed by Inserts.
Default is BA, which processes Updates as Deletes followed by
Inserts.
For more information, see “Configuring Image Type” on page 117 .

Application No Both The application name.

Name

RestartToken Yes Both The name of folder that contains the restart token folder.
File Folder Default is $PMRootDir/Restart.

RestartToken No Both The name of the restart token file.

File Name

Number of No Both The maximum number of backup copies to keep of the restart
Runs to Keep token file.
RestartToken Default is 0.
File

Recovery No Both The name of the file cache folder that is used for recovery of the
Cache Folder session.
Default is $PMRootDir/Cache.

UOW Count No Both The number of units of work (UOWs) that PWXPC processes
before it sends a commit request to PowerCenter to commit data to
all targets in the CDC session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before
sending a commit request.
Default is 1.
For more information, see “Configuring UOW Count” on page 121 .

Reader Time No Real The number of seconds that PowerCenter Integration Service
Limit Time reads data from the source before stopping.
Default is 0, which indicates that the reader time is not limited. This
value is intended for testing purposes only.
Tip: Use Idle Time instead.

NRDB CDC Application Connections 85

 Connection Required? Change Description
Attribute or Real
Time

Idle Time No Real The number of seconds the PowerExchange Listener remains idle
Time after reaching the end of the change log, as indicated by message
PWX-09967, before returning an end-of-file (EOF).
Valid values:
-1. The EOF is never returned. The session runs continuously.
0. The EOF is returned at the end of log. Then the session
terminates successfully.
n. The number of seconds that the PowerExchange Listener
remains idle before returning an EOF.
Default is -1.
For more information, see “Configuring Idle Time” on page 119 .

Real-Time No Real The number of milliseconds that must elapse before PWXPC sends
Flush Latency Time a commit request to PowerCenter. When this period elapses,
in milli- PWXPC continues to read the changes in the current UOW until it
seconds reaches the end of the UOW. Then PWXPC sends a commit
request to PowerCenter so that the data can be committed to the
targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse
before PWXPC sends a commit request to PowerCenter. If you
specify a value from 0 through 2000, PWXPC uses 2000
milliseconds (2 seconds).
Default is 0.
For more information, see “Configuring PWX Latency and Real-Time
Flush Latency” on page 122 .

PWX Latency No Real The maximum time, in seconds, that PowerExchange on the
in seconds Time source platform waits for more change data before it flushes to
PWXPC on the PowerCenter Integration Service platform.
Default is 2.

Maximum No Both The maximum number of change records that PWXPC processes
Rows Per before it sends a commit request to PowerCenter to commit data to
commit all targets in the CDC session. PWXPC does not wait for a UOW
boundary to commit the change data.
Default is 0, which disables this attribute.
For more information, see “Configuring Maximum Rows per
Commit” on page 124 .

Minimum Rows No Real The minimum number of change records that the PowerExchange
Per commit Time Listener must read from the change stream before it passes a
commit record to PWXPC. Before reaching this minimum, the
PowerExchange Listener passes only change records to PWXPC
and discards any commit records.
Default is 0, which disables this attribute.
For more information, see “Configuring Minimum Rows per
Commit” on page 125 .

86 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

Convert No Both Controls whether to convert character fields to string fields so that
character data embedded nulls in data are processed as spaces.
to string By default, this attribute is not selected.

Event Table No Real The PowerExchange extraction map name that is used for event
Time processing.
For more information, see “Configuring Event Table Processing” on
page 118 .

Offload No Real Controls whether to use CDC offload processing to move

Processing Time PowerExchange processing for the change data from the source
system to the PowerCenter Integration Service machine.
Valid values:
- No
- Yes
- Auto. PowerExchange determines whether to use offload
processing.
Default is No.

Worker No Real If you enable CDC offload processing, the number of threads that
Threads Time PowerExchange uses on the PowerCenter Integration Service
machine to process change data.
For optimal performance, the value should not exceed the number
of installed or available processors on the PowerCenter Integration
Service machine.
Valid values are from 1 through 64.
Default is 0.

Array Size No Real If the Worker Threads value is greater than 0, the size of the
Time storage array, in number of records, for the threads.
Valid values are from 1 through 100000.
Default is 25.

Map Location No Real If you enable offload processing, the location of the source
Time extraction maps. This value must be the name of a location in a
NODE statement in the DBMOVER configuration file on the
PowerCenter Integration Service machine.

Map Location No Real If you specified Map Location, a valid user ID for the user of that
User Time location.

Map Location No Real If you specify Map Location User, a valid password for that user
Password Time ID.

CAPI No Real A name that overrides the default CAPI connection name.
Connection Time
Name Override

NRDB CDC Application Connections 87

 Connection Required? Change Description
Attribute or Real
Time

Retrieve PWX No Both Controls whether PWXPC writes PowerExchange informational

Log Entries messages related to a CDC session in the PowerCenter session
log.
Default is to write PowerExchange error and warning messages but
not informational messages to the session log.

88 Chapter 4: Connections
 Connection Required? Change Description
Attribute or Real
Time

PWX Override No Both PowerExchange connection override values, separated by a

semicolon. You can enter one or more of the following overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the
application buffer for the connection. Overrides the APPBUFSIZE
statement in the DBMOVER configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the
characters in the connection string. Enter this override if
PowerExchange issues message PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves
hexadecimal '0' values, called low values, in source character fields
when passing these values to a PowerCenter session. Enter Y to
preserve low values. When the session runs, PWXPC can write
these values to a VSAM target on z/OS or to a sequential file target
on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES
statement in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements
that PWXPC uses to select data for session processing, PWXPC
adds double-quotation marks (") around every table name and field
name that includes spaces or special characters. These special
characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only
around the "table_name" value.
If for any reason, you cannot use quoted table names or field names
in your environment, set PWXNOQUOTES to Y to override the
default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes
PowerExchange error and warning messages but not informational
messages to the session log. If you specify RETLOGINFOMSG=Y,
PWXPC writes informational messages as well as error and warning
messages to the session log.
Note: The RETLOGINFOMSG connection override is equivalent to
the Retrieve PWX Log Entries connection attribute.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network
operation timeout interval in seconds. For more information about
this timeout override, see “Asynchronous Network
Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout
interval in seconds. For more information about this timeout
override, see “Asynchronous Network Communication” on page
130.

Connection No Both The number of seconds the PowerCenter Integration Service

Retry Period attempts to reconnect to the PowerExchange Listener after the
initial connection attempt fails. If the Integration Service cannot
connect to the PowerExchange Listener within the retry period, the
session fails.
Default value is 0, which disables connection retries based on this
PWXPC connection attribute.
Note: PowerExchange Listener connections to source databases
do not have connection resilience.

For more information about CDC offload processing, see the PowerExchange CDC Guide for z/OS.

NRDB CDC Application Connections 89

 Related Topics:
• “DB2 CDC Application Connections” on page 72
• “DB2 Relational Connections” on page 65
• “Microsoft SQL Server CDC Application Connections” on page 96
• “Microsoft SQL Server Relational Connections” on page 93
• “NRDB Batch Application Connections” on page 79
• “NRDB Lookup Relational Connections” on page 90
• “Oracle CDC Application Connections” on page 105
• “Oracle Relational Connections” on page 102
• “Configuring PowerExchange Connections” on page 64
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Retrieving PowerExchange Log Entries” on page 115
• “Configuring Image Type” on page 117
• “Configuring CDC Restart Attributes” on page 121
• “Configuring UOW Count” on page 121
• “Configuring Idle Time” on page 119
• “Configuring the CAPI Connection Name Override” on page 119
• “Configuring PWX Latency and Real-Time Flush Latency” on page 122
• “Configuring Maximum Rows per Commit” on page 124
• “Commit Processing for CDC Sessions” on page 161
• “Managing Session Log and Restart Token File History” on page 185
• “Configuring Event Table Processing” on page 118
• “Enabling Session Recovery for a CDC Session” on page 177

NRDB Lookup Relational Connections

Configure NRDB Lookup relational connections if you use lookups to extract data from PowerExchange
nonrelational sources.

The following table describes the connection attributes for the PWX NRDB Lookup relational connection type:

Connection Required Description

Attribute

Name Yes Name of the relational connection.

Code Page Yes Select the code page for the PowerCenter Integration Service to use to extract
data from the source database.
Note: In Unicode mode, PWXPC sets the code page with this value overriding
any code page specification in the PowerExchange configuration file.

90 Chapter 4: Connections
Connection Required Description
Attribute

Location Yes Location of the source or target database as specified as a node in the
PowerExchange configuration file dbmover.cfg.

User Name Yes User name for the PowerExchange Listener to which you are connecting.
For connections to a PowerExchange Listener on a supported Linux, UNIX, or
Windows system, if you have enabled PowerExchange LDAP user authentication,
the user name is the enterprise user name. For more information, see the
PowerExchange Reference Manual.

Password Yes A password for the specified user or a valid PowerExchange passphrase.
Passwords that are less than nine characters in length are not validated.
To avoid errors that might arise due to code page differences, Informatica
recommends that the password includes only those characters that are allowed
for passphrases.
A PowerExchange passphrase for z/OS access can be from 9 to 128 characters
in length when you use a PWXPC connection. A passphrase can contain the
following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation marks
(“), or currency symbols.
The allowable characters in the IBM IRRPHREX exit do not affect the allowable
characters in PowerExchange passphrases.
Note: A valid RACF passphrase can be up to 100 characters in length.
PowerExchange truncates passphrases longer than 100 characters when passing
them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member. For
more information, see "SECURITY Statement" in the PowerExchange Reference
Manual.
To use passphrases for lookup connections to IMS, ensure that the following
additional requirements are met:
- You must configure ODBA access to IMS as described in the PowerExchange
Navigator User Guide.
- You must use IMS data maps that specify IMS ODBA as the access method. Do not
use data maps that specify the DL/1 BATCH access method because this access
method requires the use of netport jobs, which do not support passphrases.
- The IMS database must be online in the IMS control region to use ODBA access to
IMS.

Environment No SQL commands run in the database environment.

SQL

Compression No Select to compress source data during the PowerCenter session.

Encryption No The encryption type. For information about supported encryption types, see
Type “Configuring Encryption and Compression” on page 112 .
Default is None.

NRDB Lookup Relational Connections 91

 Connection Required Description
Attribute

Encryption No The encryption level. For information about supported encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size No Enter the pacing size.

Interpret as No Specifies whether, or not, pacing size is in number of rows.

Rows

Convert No Convert character fields to string fields so that embedded nulls in data are
character data processed as spaces.
to string By default, this attribute is not selected.

PWX Override No Specifies PowerExchange connection override values, separated by semicolons,

including the following overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the application buffer for
the connection. Overrides the APPBUFSIZE statement in the DBMOVER
configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the characters in the
connection string. Enter this override if PowerExchange issues message
PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves hexadecimal '0'
values, called low values, in source character fields when passing these values to a
PowerCenter session. Enter Y to preserve low values. When the session runs,
PWXPC can write these values to a VSAM target on z/OS or to a sequential file
target on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES statement in
the DBMOVER configuration file.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes PowerExchange error and
warning messages but not informational messages to the session log. If you specify
RETLOGINFOMSG=Y, PWXPC writes informational messages as well as error and
warning messages to the session log.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network operation timeout
interval in seconds. For more information about this timeout override, see
“Asynchronous Network Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout interval in
seconds. For more information about this timeout override, see “Asynchronous
Network Communication” on page 130.

Connection No The number of seconds the PowerCenter Integration Service attempts to

Retry Period reconnect to the PowerExchange Listener after the initial connection attempt
fails. If the Integration Service cannot connect to the PowerExchange Listener
within the retry period, the session fails.
Default value is 0, which disables connection retries based on this PWXPC
connection attribute.
Note: PowerExchange Listener connections to source databases do not have
connection resilience.

Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114

92 Chapter 4: Connections
Microsoft SQL Server Relational Connections
Configure relational connections to extract data from and load data to Microsoft SQL Server sources and
targets in bulk data movement sessions. Also use relational connections to apply change data to Microsoft
SQL Server relational targets in CDC sessions.

The following table describes the connection attributes for the PWX MSSQLServer relational connection type:

Connection Required Description

Attribute

Name Yes The name of the relational connection.

Code Page Yes The code page that the PowerCenter Integration Service uses to extract data
from the source database.
Note: In Unicode mode, PWXPC uses this code page instead of any code page
specification in the PowerExchange dbmover.cfg configuration file.

Location Yes The location of the source or target database as specified in a NODE statement
in the PowerExchange dbmover.cfg configuration file.

Server Name Yes The name of the SQL Server instance.

Database Name Yes The name of the SQL Server database.

User Name Yes A user name that can be used to access the database.

Password Yes The password for the specified user name.

Environment No SQL commands that run in the database environment.

SQL

Compression No Select this option to compress source data during the PowerCenter session.

Encryption Type No The encryption type. For information about supported encryption types, see
“Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No The encryption level. For information about supported encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size No The pacing size. For information about setting the pacing size, see “Configuring
Pacing” on page 113 .
Default is 0.

Interpret as No Controls whether the pacing size is in number of rows.

Rows

Convert No Controls whether character fields are converted to string fields so that
character data embedded nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on page 114 .

Microsoft SQL Server Relational Connections 93

 Connection Required Description
Attribute

Write Mode No The write mode. For information about Write Mode options, see “Configuring
Write Mode” on page 115 .
Default is Confirm Write On.

Array Size No The size of the storage array, in number of records, for SQL Server bulk load
operations.
Valid values are from 1 through 100000. Default is 25.

Reject File No Overrides the default prefix of PWXR for the reject file.
PowerExchange creates the reject file on the target machine when Write Mode
is set to Asynchronous with Fault Tolerance.
Note: Enter PWXDISABLE to prevent creation of the reject files.

94 Chapter 4: Connections
 Connection Required Description
Attribute

PWX Override No PowerExchange connection overrides, separated by a semicolon. You can enter
one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the application buffer for
the connection. Overrides the APPBUFSIZE statement in the DBMOVER
configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the characters in the
connection string. Enter this override if PowerExchange issues message
PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves hexadecimal '0'
values, called low values, in source character fields when passing these values to
a PowerCenter session. Enter Y to preserve low values. When the session runs,
PWXPC can write these values to a VSAM target on z/OS or to a sequential file
target on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES statement
in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements that PWXPC
uses to select data for session processing, PWXPC adds double-quotation marks
(") around every table name and field name that includes spaces or special
characters. These special characters are /+-=~`!%^&*()[]{}';?,< >\\|.
If a table is identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only around the
"table_name" value.
If for any reason, you cannot use quoted table names or field names in your
environment, set PWXNOQUOTES to Y to override the default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes PowerExchange error and
warning messages but not informational messages to the session log. If you
specify RETLOGINFOMSG=Y, PWXPC writes informational messages as well as
error and warning messages to the session log.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network operation timeout
interval in seconds. For more information about this timeout override, see
“Asynchronous Network Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout interval in
seconds. For more information about this timeout override, see “Asynchronous
Network Communication” on page 130.

Connection No The number of seconds the PowerCenter Integration Service attempts to

Retry Period reconnect to the PowerExchange Listener after the initial connection attempt
fails. If the Integration Service cannot connect to the PowerExchange Listener
within the retry period, the session fails.
Default value is 0, which disables connection retries based on this PWXPC
connection attribute.
Note: PowerExchange Listener connections to source databases do not have
connection resilience.

Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Configuring Write Mode” on page 115

Microsoft SQL Server Relational Connections 95

Microsoft SQL Server CDC Application Connections
Configure CDC application connections to extract change data from Microsoft SQL Server sources in
real-time extraction mode or from PowerExchange Logger for Linux, UNIX, and Windows log files in batch or
continuous extraction mode.

The following table describes the connection attributes for the MSSQL CDC Change and CDC Real Time
application connection types:

Connection Required? Change Description

Attribute or Real
Time

Name Yes Both The name of the application connection.

Code Page Yes Both The code page that the PowerCenter Integration Service uses to
extract data from the source database.
Note: In Unicode mode, PWXPC uses this code page instead of
any code page specification in the PowerExchange dbmover.cfg
configuration file.

Location Yes Both The location of the source or target database as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file.

User Name Yes Both A user name that can be used to access the SQL Server database.

Password Yes Both The password for the specified user name.

Compression No Both Select this option to compress source data during the PowerCenter
session.

Encryption Yes Both The encryption type. For information about encryption types, see
Type “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No Both The encryption level. For information about encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size Yes Both The pacing size. For information about setting the pacing size, see
“Configuring Pacing” on page 113 .
Default is 0.

Interpret as No Both Controls whether the pacing size is in number of rows.

Rows

Image Type No Both Indicates whether PWXPC passes captured Update operations to
CDC sessions as Updates, or as Deletes followed by Inserts.
Default is BA, which processes Updates as Deletes followed by
Inserts.
For more information, see “Configuring Image Type” on page 117 .

Application No Both The application name.

Name

96 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

RestartToken Yes Both The restart token folder.

File Folder Default is $PMRootDir/Restart.
For more information, see “Configuring CDC Restart Attributes” on
page 121 .

RestartToken No Both The restart token file name.

File Name

Number of No Both The maximum number of backup copies of the restart token file to
Runs to Keep keep.
RestartToken Default is 0.
File

Recovery No Both The file cache folder to enable recovery for the session.
Cache Folder Default is $PMRootDir/Cache.

UOW Count No Both The number of units of work (UOWs) that PWXPC processes
before it sends a commit request to PowerCenter to commit data to
all targets in the CDC session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before
sending a commit request.
Default is 1.
For more information, see “Configuring UOW Count” on page 121 .

Reader Time No Real The number of seconds that the PowerCenter Integration Service
Limit Time reads data from the source before stopping.
If you enter 0, the Reader Time Limit attribute does not limit the
reader time. This attribute is intended for testing purposes only.
Tip: You can use Idle Time instead of Reader Time Limit.
Default is 0.

Idle Time No Real The number of seconds that the PowerExchange Listener remains
Time idle after reaching the end of the change log, as indicated by
message PWX-09967, before returning an end-of-file (EOF).
Valid values:
-1. The EOF is not returned. The session runs continuously.
0. The EOF is returned at the end of log. The session terminates
successfully.
n . The number of seconds after which the EOF is returned.
Default is -1.
For more information, see “Configuring Idle Time” on page 119 .

Microsoft SQL Server CDC Application Connections 97

 Connection Required? Change Description
Attribute or Real
Time

Real-Time No Real The number of milliseconds that must elapse before PWXPC sends
Flush Latency Time a commit request to PowerCenter. When this period elapses,
in milli- PWXPC continues to read the changes in the current UOW until it
seconds reaches the end of the UOW. Then PWXPC sends a commit
request to PowerCenter so that the data can be committed to the
targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse
before PWXPC sends a commit request to PowerCenter. If you
specify a value from 0 through 2000, PWXPC uses 2000
milliseconds (2 seconds).
Default is 0.
For more information, see “Configuring PWX Latency and Real-Time
Flush Latency” on page 122 .

PWX Latency No Real The maximum number of seconds that PowerExchange extraction
in seconds Time processing on the source platform waits for more change data
before it returns control to PWXPC on the PowerCenter Integration
Service machine.
Default is 2.

Maximum No Both The maximum number of change records that PWXPC processes
Rows Per before it sends a commit request to PowerCenter to commit data to
commit all targets in the CDC session. PWXPC does not wait for a UOW
boundary to commit the change data.
Default is 0, which disables this attribute.
For more information, see “Configuring Maximum Rows per
Commit” on page 124 .

Minimum Rows No Real The minimum number of change records that the PowerExchange
Per commit Time Listener must read from the change stream before it passes a
commit record to PWXPC. Before reaching this minimum, the
PowerExchange Listener passes only change records to PWXPC
and discards any commit records.
Default is 0, which disables this attribute.
For more information, see “Configuring Minimum Rows per
Commit” on page 125 .

Convert No Both Controls whether to convert character fields to string fields so that
character data embedded nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on
page 114 .

Event Table No Real The PowerExchange extraction map name that is used for event
Time processing.
For more information, see “Configuring Event Table Processing” on
page 118 .

98 Chapter 4: Connections
Connection Required? Change Description
Attribute or Real
Time

Offload No Real Controls whether PowerExchange uses CDC offload processing.

Processing Time When offload processing is enabled, PowerExchange transfers
column-level processing of the change data from the source
system to the PowerCenter Integration Service machine.
Valid values:
- No. Disables offload processing.
- Yes. Enables offload processing.
- Auto. PowerExchange determines whether to use offload
processing.
Default is No.
For more information about offload processing, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Worker No Real If you use CDC offload processing, the number of threads that
Threads Time PowerExchange uses on the PowerCenter Integration Service
machine to process change data.
For optimal performance, this value should not exceed the number
of installed or available processors on the Integration Service
machine.
Valid values are from 1 through 64.
Default is 0.

Array Size No Real If the Worker Threads value is greater than 0, specifies the size of
Time the storage array, in number of records, for the threads.
Valid values are from 1 through 100000.
Default is 25.

Map Location No Real If you use remote logging to a PowerExchange Logger for Linux,
Time UNIX, and Windows, the location of the source extraction maps.
This value must be the node name of the location, as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file on the Integration Service machine.
For more information about remote logging, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Map Location No Real If you specify a Map Location value, a valid user ID for the map
User Time location.

Map Location No Real If you specify a Map Location value, a valid password for the
Password Time specified map location user.

CAPI No Real Overrides the default CAPI connection name.

Connection Time For more information, see “Configuring the CAPI Connection Name
Name Override Override” on page 119 .

Logger DBID No Real The DBID parameter value that is specified in the PowerExchange
Time Logger for Linux, UNIX, and Windows configuration file, pwxccl.
This value is required only if the PowerExchange Logger extracts
change data for articles in multiple publication databases. In this
case, you must also set the MULTIPUB parameter to Y in the
MSQL CAPI_CONNECTION statement in the PowerExchange
dbmover.cfg configuration file. Otherwise, the extraction fails.

Microsoft SQL Server CDC Application Connections 99

 Connection Required? Change Description
Attribute or Real
Time

Retrieve PWX No Both Controls whether PWXPC writes PowerExchange informational

Log Entries messages related to a CDC session in the PowerCenter session
log.
Default is to write PowerExchange error and warning messages but
not informational messages to the session log.

100 Chapter 4: Connections

Connection Required? Change Description
Attribute or Real
Time

PWX Override No Both PowerExchange connection overrides, separated by a semicolon.

You can enter one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the
application buffer for the connection. Overrides the value of
APPBUFSIZE in the DBMOVER configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the
characters in the connection string. Enter this override if
PowerExchange issues message PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves
hexadecimal '0' values, called low values, in source character fields
when passing these values to a PowerCenter session. Enter Y to
preserve low values. When the session runs, PWXPC can write
these values to a VSAM target on z/OS or to a sequential file target
on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES
statement in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements
that PWXPC uses to select data for session processing, PWXPC
adds double-quotation marks (") around every table name and field
name that includes spaces or special characters. These special
characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only
around the "table_name" value.
If for any reason, you cannot use quoted table names or field names
in your environment, set PWXNOQUOTES to Y to override the
default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes
PowerExchange error and warning messages but not informational
messages to the session log. If you specify RETLOGINFOMSG=Y,
PWXPC writes informational messages as well as error and warning
messages to the session log.
Note: The RETLOGINFOMSG connection override is equivalent to
the Retrieve PWX Log Entries connection attribute.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network
operation timeout interval in seconds. For more information about
this timeout override, see “Asynchronous Network
Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout
interval in seconds. For more information about this timeout
override, see “Asynchronous Network Communication” on page
130.

Connection No Both The number of seconds the PowerCenter Integration Service

Retry Period attempts to reconnect to the PowerExchange Listener after the
initial connection attempt fails. If the Integration Service cannot
connect to the PowerExchange Listener within the retry period, the
session fails.
Default value is 0, which disables connection retries based on this
PWXPC connection attribute.
Note: PowerExchange Listener connections to source databases
do not have connection resilience.

Microsoft SQL Server CDC Application Connections 101

 Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Retrieving PowerExchange Log Entries” on page 115
• “Configuring Image Type” on page 117
• “Configuring Idle Time” on page 119
• “Configuring the CAPI Connection Name Override” on page 119
• “Configuring UOW Count” on page 121
• “Configuring CDC Restart Attributes” on page 121
• “Configuring PWX Latency and Real-Time Flush Latency” on page 122
• “Configuring Maximum Rows per Commit” on page 124
• “Commit Processing for CDC Sessions” on page 161
• “Managing Session Log and Restart Token File History” on page 185
• “Enabling Session Recovery for a CDC Session” on page 177

Oracle Relational Connections

Configure relational connections to extract data from and load data to Oracle sources and targets in bulk data
movement sessions. Also use relational connections to apply change data to Oracle relational targets in CDC
sessions.

The following table describes the connection attributes for the PWX Oracle relational connection type:

Connection Required Description

Attribute

Name Yes The name of the relational connection.

Code Page Yes The code page that the PowerCenter Integration Service uses to extract data
from the source database.
Note: In Unicode mode, PWXPC uses this code page instead of any code page
specification in the PowerExchange dbmover.cfg configuration file.

Location Yes The location of the source or target database as specified in a NODE statement
in the PowerExchange dbmover.cfg configuration file.

TNS Name Yes The Net Service Name of the Oracle instance.

User Name Yes A user name that can be used to access the database.
For databases on supported Linux, UNIX, or Windows systems, if you have
enabled PowerExchange LDAP user authentication and disabled relational pass-
through authentication, the user name is the enterprise user name. For more
information, see the PowerExchange Reference Manual.

Password Yes The password for the specified user name.

102 Chapter 4: Connections

Connection Required Description
Attribute

Environment No SQL commands that run in the database environment.

SQL

Compression No Select this option to compress source data during the PowerCenter session.

Encryption No The encryption type. For information about supported encryption types, see
Type “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No The encryption level. For information about supported encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .

Pacing Size No The pacing size. For information about setting the pacing size, see “Configuring
Pacing” on page 113 .
Default is 0.

Interpret as No Controls whether the pacing size is in number of rows.

Rows

Convert No Controls whether character fields are converted to string fields so that embedded
character data nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on page 114 .

Write Mode No The write mode. For information about Write Mode options, see “Configuring
Write Mode” on page 115 .
Default is Confirm Write On.

Reject File No Overrides the default prefix of PWXR for the reject file.
PowerExchange creates the reject file on the target machine when Write Mode
is set to Asynchronous with Fault Tolerance.
Note: Enter PWXDISABLE to prevent creation of the reject files.

Oracle Relational Connections 103

 Connection Required Description
Attribute

PWX Override No PowerExchange connection overrides, separated by a semicolon. You can enter
one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the application buffer for
the connection. Overrides the APPBUFSIZE statement in the DBMOVER
configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the characters in the
connection string. Enter this override if PowerExchange issues message
PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves hexadecimal '0'
values, called low values, in source character fields when passing these values to
a PowerCenter session. Enter Y to preserve low values. When the session runs,
PWXPC can write these values to a VSAM target on z/OS or to a sequential file
target on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES statement in
the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements that PWXPC
uses to select data for session processing, PWXPC adds double-quotation marks
(") around every table name and field name that includes spaces or special
characters. These special characters are /+-=~`!%^&*()[]{}';?,< >\\|.
If a table is identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only around the
"table_name" value.
If for any reason, you cannot use quoted table names or field names in your
environment, set PWXNOQUOTES to Y to override the default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes PowerExchange error and
warning messages but not informational messages to the session log. If you
specify RETLOGINFOMSG=Y, PWXPC writes informational messages as well as
error and warning messages to the session log.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network operation timeout
interval in seconds. For more information about this timeout override, see
“Asynchronous Network Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout interval in
seconds. For more information about this timeout override, see “Asynchronous
Network Communication” on page 130.

Connection No The number of seconds the PowerCenter Integration Service attempts to

Retry Period reconnect to the PowerExchange Listener after the initial connection attempt
fails. If the Integration Service cannot connect to the PowerExchange Listener
within the retry period, the session fails.
Default value is 0, which disables connection retries based on this PWXPC
connection attribute.
Note: PowerExchange Listener connections to source databases do not have
connection resilience.

For more information, see the PowerExchange Bulk Data Movement Guide.

Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Converting Character Data to Strings” on page 114
• “Configuring Write Mode” on page 115

104 Chapter 4: Connections

Oracle CDC Application Connections
Configure CDC application connections to extract change data from Oracle sources in real-time extraction
mode or from PowerExchange Logger for Linux, UNIX, and Windows log files in batch or continuous
extraction mode.

The following table describes the connection attributes for the PWX Oracle CDC application connection type:

Connection Required? Change Description

Attribute or Real
Time

Name Yes Both The name of the application connection.

Code Page Yes Both The code page that the PowerCenter Integration Service uses to
extract data from the source database.
In Unicode mode, you must select UTF-8 for Oracle CDC
connections.
Note: In Unicode mode, PWXPC sets the code page with this value
overriding any code page specification in the PowerExchange
configuration file.

Location Yes Both The location of the source or target database as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file.

User Name Yes Both A user name that can be used to access the Oracle database.
For a database on a supported Linux, UNIX, or Windows system, if
you have enabled PowerExchange LDAP user authentication, the
user name is the enterprise user name. For more information, see
the PowerExchange Reference Manual.

Password Yes Both The password for the specified user name.

Compression No Both Select this option to compress source data during the PowerCenter
session.

Encryption Yes Both The encryption type. For information about encryption types, see
Type “Configuring Encryption and Compression” on page 112 .
Default is None.

Encryption No Both The encryption level. For information about encryption levels, see
Level “Configuring Encryption and Compression” on page 112 .
Default is 1.

Pacing Size Yes Both The pacing size. For information about setting the pacing size, see
“Configuring Pacing” on page 113 .
Default is 0.

Interpret as No Both Controls whether the pacing size is in number of rows.

Rows

Oracle CDC Application Connections 105

 Connection Required? Change Description
Attribute or Real
Time

Image Type No Both Indicates whether PWXPC passes captured Update operations to
CDC sessions as Updates, or as Deletes followed by Inserts.
Default is BA, which processes Updates as Deletes followed by
Inserts.
For more information, see “Configuring Image Type” on page 117 .

Application No Both The application name.

Name

RestartToken Yes Both The restart token folder.

File Folder Default is $PMRootDir/Restart.
For more information, see “Configuring CDC Restart Attributes” on
page 121 .

RestartToken No Both The restart token file name.

File Name

Number of No Both The maximum number of backup copies of the restart token file to
Runs to Keep keep.
RestartToken Default is 0.
File

Recovery No Both The file cache folder to enable recovery for the session.
Cache Folder Default is $PMRootDir/Cache.

UOW Count No Both The number of units of work (UOWs) that PWXPC processes
before it sends a commit request to PowerCenter to commit data to
all targets in the CDC session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before
sending a commit request.
Default is 1.
For more information, see “Configuring UOW Count” on page 121 .

Reader Time No Real The number of seconds that the PowerCenter Integration Service
Limit Time reads data from the source before stopping.
If you enter 0, the Reader Time Limit attribute does not limit the
reader time. This attribute is intended for testing purposes only.
Tip: Use Idle Time instead of Reader Time Limit.
Default is 0.

106 Chapter 4: Connections

Connection Required? Change Description
Attribute or Real
Time

Idle Time No Real The number of seconds that the PowerExchange Listener remains
Time idle after reaching the end of the change log, as indicated by
message PWX-09967, before returning an end-of-file (EOF).
Valid values:
-1. The EOF is not returned. The session runs continuously.
0. The EOF is returned at the end of log. The session terminates
successfully.
n . The number of seconds after which the EOF is returned.
Default is -1.
For more information, see “Configuring Idle Time” on page 119 .

Real-Time No Real The number of milliseconds that must elapse before PWXPC sends
Flush Latency Time a commit request to PowerCenter. When this period elapses,
in milli- PWXPC continues to read the changes in the current UOW until it
seconds reaches the end of the UOW. Then PWXPC sends a commit
request to PowerCenter so that the data can be committed to the
targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse
before PWXPC sends a commit request to PowerCenter. If you
specify a value from 0 through 2000, PWXPC uses 2000
milliseconds (2 seconds).
Default is 0.
For more information, see “Configuring PWX Latency and Real-Time
Flush Latency” on page 122 .

PWX Latency No Real The maximum number of seconds that PowerExchange extraction
in seconds Time processing on the source platform waits for more change data
before it returns control to PWXPC on the PowerCenter Integration
Service machine.
Default is 2.

Maximum No Both The maximum number of change records that PWXPC processes
Rows Per before it sends a commit request to PowerCenter to commit data to
commit all targets in the CDC session. PWXPC does not wait for a UOW
boundary to commit the change data.
Default is 0, which disables this attribute.
For more information, see “Configuring Maximum Rows per
Commit” on page 124 .

Minimum Rows No Real The minimum number of change records that the PowerExchange
Per commit Time Listener must read from the change stream before it passes a
commit record to PWXPC. Before reaching this minimum, the
PowerExchange Listener passes only change records to PWXPC
and discards any commit records.
Default is 0, which disables this attribute.
For more information, see “Configuring Minimum Rows per
Commit” on page 125 .

Oracle CDC Application Connections 107

 Connection Required? Change Description
Attribute or Real
Time

Instance Name No Real Overrides the Oracle instance name in the ORACLEID statement in
Time the PowerExchange dbmover.cfg configuration file.

Connect String No Real Overrides the connection value in the ORACLEID statement in the
Time PowerExchange dbmover.cfg configuration file.

Source No Both Overrides the schema name for a group of registrations.

Schema Name

Convert No Both Controls whether to convert character fields to string fields so that
character data embedded nulls in data are processed as spaces.
to string By default, this attribute is not selected.
For more information, see “Converting Character Data to Strings” on
page 114 .

Event Table No Real The PowerExchange extraction map name that is used for event
Time processing.
For more information, see “Configuring Event Table Processing” on
page 118 .

Offload No Real Controls whether PowerExchange uses CDC offload processing.

Processing Time When offload processing is enabled, PowerExchange transfers
column-level processing of the change data from the source
system to the PowerCenter Integration Service machine.
Valid values:
- No. Disables offload processing.
- Yes. Enables offload processing.
- Auto. PowerExchange determines whether to use offload
processing.
Default is No.
For more information about offload processing, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Worker No Real If you use CDC offload processing, the number of threads that
Threads Time PowerExchange uses on the PowerCenter Integration Service
machine to process change data.
For optimal performance, this value should not exceed the number
of installed or available processors on the PowerCenter Integration
Service machine.
Valid values are from 1 through 64.
Default is 0.

Array Size No Real If the Worker Threads value is greater than 0, specifies the size of
Time the storage array, in number of records, for the threads.
Valid values are from 1 through 100000.
Default is 25.

108 Chapter 4: Connections

Connection Required? Change Description
Attribute or Real
Time

Map Location No Real If you use remote logging to a PowerExchange Logger for Linux,
Time UNIX, and Windows, the location of the source extraction maps.
This value must be the node name of the location, as specified in a
NODE statement in the PowerExchange dbmover.cfg configuration
file on the Integration Service machine.
For more information about remote logging, see the
PowerExchange CDC Guide for Linux, UNIX, and Windows.

Map Location No Real If you specify a Map Location value, a valid user ID for the map
User Time location.

Map Location No Real If you specify a Map Location value, a valid password for the
Password Time specified map location user.

CAPI No Real Overrides the default CAPI connection name.

Connection Time
Name Override

Retrieve PWX No Both Controls whether PWXPC writes PowerExchange informational

Log Entries messages related to a CDC session in the PowerCenter session
log.
Default is to write PowerExchange error and warning messages but
not informational messages to the session log.

Oracle CDC Application Connections 109

 Connection Required? Change Description
Attribute or Real
Time

PWX Override No Both PowerExchange connection overrides, separated by a semicolon.

You can enter one or more of the following optional overrides:
- APPBUFSIZE=app_buf_size. Specifies the initial size of the
application buffer for the connection. Overrides the APPBUFSIZE
statement in the DBMOVER configuration file.
- CONNECTSTRINGCODEPAGE=code_page. Code page of the
characters in the connection string. Enter this override if
PowerExchange issues message PWX-07610.
- LOWVALUES={Y|N}. Indicates whether PowerExchange preserves
hexadecimal '0' values, called low values, in source character fields
when passing these values to a PowerCenter session. Enter Y to
preserve low values. When the session runs, PWXPC can write
these values to a VSAM target on z/OS or to a sequential file target
on Linux, UNIX, Windows, or z/OS. Overrides the LOWVALUES
statement in the DBMOVER configuration file.
- PWXNOQUOTES={Y|N}. By default, in the SQL SELECT statements
that PWXPC uses to select data for session processing, PWXPC
adds double-quotation marks (") around every table name and field
name that includes spaces or special characters. These special
characters are /+-=~`!%^&*()[]{}';?,< >\\|. If a table is
identified by both a schema name and table name in the format
schema.table_name, PWXPC places the quotation marks only
around the "table_name" value.
If for any reason, you cannot use quoted table names or field names
in your environment, set PWXNOQUOTES to Y to override the
default behavior.
- RETLOGINFOMSG={N|Y}. By default, PWXPC writes
PowerExchange error and warning messages but not informational
messages to the session log. If you specify RETLOGINFOMSG=Y,
PWXPC writes informational messages as well as error and warning
messages to the session log.
Note: The RETLOGINFOMSG connection override is equivalent to
the Retrieve PWX Log Entries connection attribute.
- TCPIP_OP_TIMEOUT=network_operation_timeout. A network
operation timeout interval in seconds. For more information about
this timeout override, see “Asynchronous Network
Communication” on page 130.
- TCPIP_CON_TIMEOUT=connection_timeout. A connection timeout
interval in seconds. For more information about this timeout
override, see “Asynchronous Network Communication” on page
130.

Connection No Both The number of seconds the PowerCenter Integration Service

Retry Period attempts to reconnect to the PowerExchange Listener after the
initial connection attempt fails. If the Integration Service cannot
connect to the PowerExchange Listener within the retry period, the
session fails.
Default value is 0, which disables connection retries based on this
PWXPC connection attribute.
Note: PowerExchange Listener connections to source databases
other than those used for PowerExchange Express CDC for Oracle
do not have connection resilience.

110 Chapter 4: Connections

 For more information about CDC offload processing and remote logging, see the PowerExchange CDC Guide
for Linux, UNIX, and Windows.

Related Topics:
• “Configuring Encryption and Compression” on page 112
• “Configuring Pacing” on page 113
• “Configuring Write Mode” on page 115
• “Converting Character Data to Strings” on page 114
• “Retrieving PowerExchange Log Entries” on page 115
• “Configuring Image Type” on page 117
• “Configuring the CAPI Connection Name Override” on page 119
• “Configuring Idle Time” on page 119
• “Configuring UOW Count” on page 121
• “Configuring CDC Restart Attributes” on page 121
• “Configuring PWX Latency and Real-Time Flush Latency” on page 122
• “Configuring Maximum Rows per Commit” on page 124
• “Commit Processing for CDC Sessions” on page 161
• “Managing Session Log and Restart Token File History” on page 185
• “Configuring Event Table Processing” on page 118
• “Enabling Session Recovery for a CDC Session” on page 177

Configuring Connection Attributes

The connection attributes that you use depend on how you want to run the PowerCenter session to extract
data from the source. This section highlights some key connection attributes. For a complete list of all
attributes that you can configure for a connection type, see the topics for specific connection types.

Use the following key connection attributes, as appropriate:

• Common connection attributes:

- Compression, Encryption Type, Encryption Level

- Pacing Size, Interpret as Rows

- Convert character data to string

- Retrieve PWX Log Entries

• Batch application and relational connection attributes:
- Write Mode
• CDC connection attributes:
- Image Type

- Event Table

- CAPI Connection Name Override

- Idle Time

Configuring Connection Attributes 111

 - CDC Restart

- UOW Count

- Real-Time Flush Latency

- Maximum Rows Per commit

- Minimum Rows Per commit

• Asynchronous communication connection attributes:
- PWX Override: TCPIP network operations timeout

- PWX Override: TCPIP connection timeout

Common Connection Attributes

The attributes discussed in this section are common to both the PWXPC relational and application connection
types.

Configuring Encryption and Compression

When you configure connections to extract or load data with PowerExchange, you can set connection
attributes to encrypt or compress data during the PowerCenter session.

Set the Compression, Encryption Type, and Encryption Level connection attributes to compress data or to
encrypt data by using PowerExchange alternative network security. These attributes override the values in
the COMPRESS, ENCRYPT, and ENCRYPTLEVEL statements, if defined, in the DBMOVER configuration
file on the Integration Service machine. To enable compression or encryption for a PowerCenter session, be
sure to select the appropriate connection attributes.

Note: Informatica recommends that you use Secure Sockets Layer (SSL) authentication instead of
configuring the Encrypt Level and Encryption Type connection attributes or the ENCRYPT and
ENCRYPTLEVEL statements in the DBMOVER configuration file. SSL authentication provides stricter
security and is used by several Informatica products.

For more information about implementing SSL authentication in a PowerExchange network, see the
PowerExchange Reference Manual.

112 Chapter 4: Connections

The following table describes the connection attributes for compression and encryption:

Connection Required Description

Attribute

Compression No Select to compress source data during the PowerCenter session.

Encryption Level No If you select RC2 or DES for Encryption Type, enter one of the following
values to indicate the encryption level:
- 1. Uses a 56-bit encryption key for both DES and RC2.
- 2. Uses a 168-bit triple DES encryption key for DES, or a 64-bit encryption
key for RC2.
- 3. Uses a 168-bit triple DES encryption key for DES, or a 128-bit encryption
key for RC2.
Default is 1.
Note: If you select None for Encryption Type, PowerExchange ignores
the Encryption Level value.

Encryption Type Yes Select one of the following values for the encryption type:
- None
- RC2
- DES
Default is None.

Configuring Pacing
You can configure the pacing size to slow the data transfer rate from the PowerExchange Listener. The
pacing size determines the amount of data the PowerExchange Listener passes to the source or target.
Configure the pacing size if an external application, database, or the PowerCenter Integration Service node is
a bottleneck during the session.

The following table describes the pacing attributes:

Connection Required Description

Attribute

Pacing Size No Enter the amount of data the source system can pass to the
PowerExchange Listener. The lower the value, the faster the session
performance.
Minimum value is 0. Enter 0 for maximum performance.
Default is 0.

Interpret as Rows No Select to represent the pacing size in number of rows. If you clear this
option, the pacing size represents kilobytes.
This option is selected by default.

Configuring Timeouts
When you configure connections to extract and load data from PowerExchange, you can configure
connection and network operation timeouts.

By default, PowerExchange uses a connection timeout value of 180 seconds, and network operation timeouts
are not enabled. Instead, PowerExchange uses heartbeat processing to detect network send and receive
timeouts.

Configuring Connection Attributes 113

 You define a connection or network operation timeout value by using the PWX Override connection attribute.

The following table describes the timeout values that you can specify with the PWX Override attribute:

Connection Required Description

Attribute

PWX Override No Specifies PowerExchange timeout values, in seconds. Separate multiple

statements with semicolons.
- TCPIP_OP_TIMEOUT= network_operation_timeout ;
- TCPIP_CON_TIMEOUT= connection_timeout
Valid values for TCPIP_OP_TIMEOUT are from 5 to 30758400. Or you can
specify 0 or -1 for no timeouts, the defaut.
Valid values for TCPIP_CON_TIMEOUT are from 10 to 2678400. Default is
180.

Converting Character Data to Strings

For some data sources, data fields that have a character datatype such as CHAR or VARCHAR might contain
embedded nulls. PowerCenter expects character fields to contain null-terminated strings. If character fields
contain embedded nulls, unpredictable results might occur for bulk data movement and CDC sessions.

Usually, this problems occurs for z/OS sources with character fields that intentionally or unintentionally
contain embedded nulls.

To avoid this problem, you can convert character fields to string fields. For string data fields, PowerCenter
converts each embedded null to a space.

For example, a single character field is a collection of subfields. The user application that writes data to the
z/OS source uses null delimiters (x’00’) to separate the subfields.

The character field contains the following hexadecimal EBCDIC data:

C1C2C300C4C5C6C700C8C9

In this data, the double zeroes represent the null delimiters. The user application parses the data into the
following hexadecimal strings:

• x’C1C2C3’ (translates to text "ABC")

• x’C4C5C6C7’ (translates to text "DEFG")
• x’C8C9’ (translates to text "HI")
No explicit mapping of the subfields exists. PowerCenter reads only the string "ABC" for the session and
truncates the rest of the data, after the first null delimiter.

Besides truncation, other unpredictable results can occur, such as session failure. Problems are less likely to
occur if the workflow passes the data directly through to the target without a transformation.

To mitigate these issues, you can convert character data with embedded nulls to strings so that the nulls are
handled as spaces. Use one of the following methods:

• To have PowerCenter convert character fields for all sources that use a connection, select the Convert
character data to string attribute on the connection. By default, this attribute is not selected and
PowerCenter passes the nulls in the character fields to the target.
• To manually convert individual character fields in a table, edit the source definition in Source Analyzer. In
the Edit Tables dialog box, click the Columns tab. Then, change the datatype for each CHAR and
VARCHAR column that you want to convert to STRING. This method is more efficient if you know which
character columns are likely to contain nulls.

114 Chapter 4: Connections

 If you apply one of these conversion methods to the preceding example EBCDIC data, PWXPC replaces the
zeroes with "40" when reading the data:
C1C2C340C4C5C6C740C8C9

When PowerCenter extracts this data to the target, it replaces the "40" values with spaces:
ABC DEFG HI

Retrieving PowerExchange Log Entries

PWXPC writes PowerExchange error and warning messages to the PowerCenter session log by default. You
can configure PWXPC to also write PowerExchange informational messages to the session log. The session
log then contains all of the PowerCenter and PowerExchange messages related to session processing.

PWXPC incorporates PowerExchange messages into PowerCenter message PWXPC_10091 when writing
messages to the session log.

To configure PWXPC to include PowerExchange informational messages in the PowerCenter session log,
configure one of the following connection attributes in PowerCenter:

• Select the Retrieve PWX Log Entries connection attribute. (PWXPC application connection types only)
• Specify the RETLOGINFOMSG=Y option in the PWX Override connection attribute. (All PWXPC
connection types)
The Retrieve PWX Log Entries connection attribute and the RETLOGINFOMSG=Y option of the PWX
Override connection attribute are equivalent.

If you enable the Retrieve PWX Log Entries connection attribute or specify RETLOGINFOMSG=Y in the
PWX Override connection attribute, PWXPC writes all PowerExchange messages, including informational
messages, that are related to the session to the PowerCenter session log.

Caution: Informational messages are not limited in size and might take up hundreds of lines in your session
log.

Batch Application and Relational Connection Attributes

These attributes are available in both PWXPC batch application and relational connection types unless
otherwise noted.

Configuring Write Mode

To configure the write mode, define the Write Mode and Reject File connection attributes.

Write Mode

Optional. Defines the mode in which to send data to the PowerExchange Listener.

Select one of the following write modes:

Confirm Write On

Sends data to the PowerExchange Listener and waits for a success or no success response before
sending more data. This mode sends data synchronously to the PowerExchange Listener rather
than buffering the data. After sending an SQL request, the sender waits for the response from
PowerExchange before sending the next SQL request.

This mode provides good error recovery but has slower data transfer rates than the other modes.

To stop session execution when the session encounters errors, specify a value greater than 0 in the
error handling option Stop on errors on the Config Object tab.

Configuring Connection Attributes 115

 For CDC workflows with a resume recovery strategy, specify a write mode of Confirm Write On in
the relational connection.

Confirm Write Off

Sends data asynchronously to the PowerExchange Listener by buffering the data. This mode does
not wait for a success or no success response. While this mode provides a faster data transfer
speed than Confirm Write On, it removes the ability to determine which SQL statement failed in
error situations. As a result, you must reload the entire table if an error occurs to ensure data
integrity. Use this setting only only if you can reload tables when an error occurs.

Note: PowerCenter statistics are unreliable when you use Confirm Write Off.

Asynchronous with Fault Tolerance

Combines the speed of Confirm Write Off with the error detection of Confirm Write On. This mode
buffers the data and sends it asynchronously to the PowerExchange Listener. PowerExchange
creates a reject file on the target machine when SQL errors occur, allowing you to correct errors
without reloading the entire table. You can also specify how to handle specific SQL return codes. To
stop session execution when the session encounters errors, specify a value greater than 0 in the
error handling option Stop on errors on the Config Object tab.

Asynchronous with Fault Tolerance is available only for PWX DB2zOS, PWX DB2i5OS, PWX
DB2LUW, PWX MSSQLServer, and PWX Oracle relational connections.

Default is Confirm Write On.

Reject File

Optional. A prefix that overrides the default prefix of PWXR for the PowerExchange reject file.

PowerExchange creates the reject file on the target machine when you specify a write mode of
Asynchronous with Fault Tolerance and an error occurs that prevents data from being written to the
target. The reject file contain the rows of data that the writer did not write to the target.

Note: Enter PWXDISABLE to prevent the creation of the reject files.

For more information, see the PowerExchange Bulk Data Movement Guide.

CDC Application Connection Attributes

This section describes attributes that are specific to the PWXPC CDC application connection types.

Related Topics:
• “Configuring Image Type” on page 117
• “Configuring Event Table Processing” on page 118
• “Configuring Idle Time” on page 119
• “Configuring Maximum Rows per Commit” on page 124
• “Configuring Minimum Rows per Commit” on page 125
• “Configuring PWX Latency and Real-Time Flush Latency” on page 122
• “Configuring the CAPI Connection Name Override” on page 119
• “Configuring UOW Count” on page 121
• “Connection Attributes for CDC Commit Processing ” on page 126
• “Extracting CDC Data in Change and Real-time Modes” on page 146

116 Chapter 4: Connections

Configuring Image Type
The following table describes the optional Image Type connection attribute:

Connection Change or Description

Attribute Real Time

Image Type Both Indicates how PWXPC passes captured updates on a source to CDC sessions
that apply the updates to the target.
Valid values:
- AI. Process updates as update operations. PWXPC passes each update as a
single update record. An update record includes after images of the data only,
unless you add before image (BI) and change indicator (CI) fields to the
extraction map that is imported to create the source definition for the CDC
session.
- BA. Process updates as deletes followed by inserts. PWXPC passes each
update as a delete record followed by an insert record. The delete record
contains the before image of the data, and the insert record contains the after
image.
Default is BA.

PowerExchange captures before and after image data for update operations on the source, regardless of the
source type. The before-image data can always be extracted in real-time mode. However, in batch and
continuous extraction modes, only after-image data is available if the changes were processed by
PowerExchange Condense or the PowerExchange Logger for Linux, UNIX, and Windows with only after
images.

If you specify BA, PWXPC generates, for each captured update operation, a delete record that contains the
before image of the data and an insert record that contains the after image. If you also define BI and CI fields
for some columns in the extraction map that is imported for the source definition, PWXPC populates the BI
and CI fields based on the information in the generated delete and insert records.

If you specify AI, you can still use before images of the data, if available, in extraction processing. PWXPC
can embed before-image data and after-image data in the same update row. For PWXPC to embed before-
image data, you must complete the following configuration tasks:

• In the PowerExchange Navigator, add BI and CI fields to the extraction map that you plan to import for the
source definition in PowerCenter.
• If you use batch or continuous extraction mode, enter BA for the CAPT_IMAGE parameter in the
PowerExchange Condense or PowerExchange Logger for Linux, UNIX, and Windows configuration file.
This setting causes both before and after images to be stored in the PowerExchange Condense condense
files or PowerExchange Logger log files. When CDC sessions run, they extract data from these files.
Informatica recommends that you use the AI setting if you want to process before images of data. CDC
sessions can process a single update record more efficiently than separate delete and insert records to get
the before image data.

For example, embed before-image data and after-image data in the same update row to handle changes to
primary keys. Relational databases that allow changes to primary keys, such as DB2 for z/OS, treat these
updates as equivalent to deleting the row and reading it with a new key value. To enable PowerExchange to
detect primary key changes, include BI and CI fields for the primary key columns in the extraction map for the
source definition. Then, in PowerCenter, define a Flexible Target Key Custom transformation to apply the
changes to the target as a delete followed by an insert. Include the transformation in the mapping for the
CDC session. If a target relational database does not allow changes to primary keys, updates to primary keys
fail.

Note: To use a Flexible Target Key Custom transformation, you must set the Image Type attribute to AI and
configure BI and CI fields in the PowerExchange extraction map for the source.

Configuring Connection Attributes 117

 For more information about adding BI and CI columns to extraction maps, see the PowerExchange Navigator
User Guide.

Related Topics:
• “Configuring Mappings with Flexible Target Key Transformations” on page 189

Configuring Event Table Processing

The following table describes the optional Event Table connection attribute:

Connection Attribute Change or Real Time Description

Event Table Real Time Specifies the PowerExchange extraction map name that is
used for event table processing.

You can use event table processing to stop the extraction of changes based on user-defined events, such as
an end-of-day event. For example, to stop an extraction process every night, after all of the changes for the
day have been processed, write a change to the event table at midnight. This change triggers
PowerExchange to stop reading change data and shut down the extraction process after the current UOW
completes.

Event table processing has the following rules and guidelines:

• You can only use event table processing with real-time or continuous extraction modes.
• You must create the event table, and define the applications that can update the table.
• You must register the event table for change data capture from the PowerExchange Navigator.
• The event table and all of the source tables in the CDC session must be of the same source type.
To use event table processing:

1. Create an event table.

The event table must be of the same source type and on the same machine as the change data that is
extracted. For example, if you extract DB2 change data on MVS, the event table must be a DB2 table in
the same DB2 subsystem as the DB2 source tables for the extraction.
2. In the PowerExchange Navigator, create a capture registration and extraction map for the event table.
When you create a capture registration, the PowerExchange Navigator generates an extraction map.
3. In PowerCenter, create a CDC session, and specify the extraction map name in the Event Table
attribute on the PWX CDC Real Time application connection.
4. When the defined event occurs, update the event table.
When PowerExchange reads the update to the event table, PowerExchange places an end-of-file (EOF)
into the change stream. PWXPC processes the EOF, passes it to the PowerCenter Integration Service,
and then shuts down the PowerExchange reader. The PowerCenter Integration Service completes
writing all of the data currently in the pipeline to the targets and then ends the CDC session.

118 Chapter 4: Connections

Configuring the CAPI Connection Name Override
The following table describes the optional CAPI Connection Name Override connection attribute:

Connection Attribute Change or Real Time Description

CAPI Connection Name Real Time Overrides the CAPI connection name from the
Override PowerExchange DBMOVER configuration file.

PowerExchange allows a maximum of eight CAPI_CONNECTION statements in a PowerExchange

DBMOVER configuration file. Define multiple CAPI_CONNECTION statements when you want to capture
changes from more than one database type through a single PowerExchange Listener on a single machine.
For example, you can capture changes for Oracle and DB2 for Linux, UNIX, and Windows sources with a
single PowerExchange Listener by specifying source-specific CAPI_CONNECTION statements. You can also
specify a default CAPI_CONNECTION statement by adding the CAPI_SRC_DFLT statement in the
DBMOVER configuration file. For more information about PowerExchange CAPI_CONNECTION statements,
see the PowerExchange Reference Manual.

To override the CAPI_CONNECTION statements in the DBMOVER configuration file for a PowerCenter CDC
session, specify the CAPI Connection Name Override connection attribute.

Configuring Idle Time

The following table describes the optional Idle Time connection attribute:

Connection Change or Description

Attribute Real Time

Idle Time Real Time Specifies the number of seconds the PowerExchange Listener remains idle
after reaching the end of the change log, as indicated by message
PWX-09967, before returning an end-of-file (EOF).
Valid values:
-1. The EOF is never returned. The session runs continuously.
0. The EOF is returned at the end of log. The session ends successfully.
n. The EOF is returned after no data is received for specified number of
seconds. The session ends.
Default is -1.

Use the Idle Time terminating condition to indicate whether a real-time session should run continuously or
shut down after a specified period of time.

The Idle Time timer starts when the PowerExchange Listener begins reading change data for the sources.

If you enter -1 for Idle Time, PowerExchange never returns an EOF to the PowerCenter Integration Server,
which causes the session to run continuously. Typically, the default value of -1 is used for real-time sessions.

You can stop continuous extraction sessions by issuing the PowerCenter Workflow Monitor Stop or Abort
command, the pmcmd commands to stop and abort tasks and workflows, or the PowerExchange STOPTASK
command.

• Stopping the session or workflow with the PowerCenter Workflow Monitor or the pmcmd stop task
commands is a normal termination. PowerCenter performs a graceful stop after the CDC reader and the
writers process all of the data in the pipeline and shutdown.

Configuring Connection Attributes 119

 • Aborting the session or workflow with the PowerCenter Workflow Monitor or the pmcmd abort task
command is an abnormal termination. PowerCenter does not wait for the CDC reader and the writers to
shut down or to process all of the data in the pipeline.
• The PowerExchange STOPTASK command stops the extraction task in the PowerExchange Listener and
passes an EOF to the Integration Service, which then ends the session successfully.

Warning: In the session properties, verify that the Commit Type property specifies Source and that the
Commit at End of File property is disabled. By default, the Commit at End of File property is enabled. This
property causes data to be committed after the CDC reader commits the restart tokens and shuts down,
which results in duplicate data being sent to the targets when the session is restarted.

If you enter 0 for Idle Time, PowerExchange returns an EOF to the PowerCenter Integration Service when
the end-of-log (EOL) is reached. After the EOF is received, the PowerCenter Integration Service terminates
the session successfully, which includes committing all of the data and updating the restart token file. The
EOL is determined by what was the current end of the change stream at the point that PowerExchange
started to read the change stream. This concept of EOL is required because the change stream is generally
not static so the actual EOL is continually moving forward. PowerExchange issues the following message
when the EOL is reached:
PWX-09967 CAPI i/f: End of log for time yy/mm/dd hh:mm:ss reached

For example, if a session starts reading a change stream at 10:00 a.m., the EOL at that point is determined.
After PowerExchange reaches that point in the change stream, it will return EOF to the Integration Service.
This means that changes recorded in the change stream after 10:00 a.m. will not be processed. Specifying 0
for Idle time is a useful in situations where you want to extract change data for sources periodically as
opposed to continuously.

If you enter a positive number for Idle Time, the session runs until no data is returned for the period of time
specified. After the Idle Time limit is reached, PowerExchange send an EOF to the PowerCenter Integration
Service and the session terminates successfully. Specifying a low value for Idle Time, such as 1, can result
in this time being reached before all available data in the change stream has been read.

The following message is issued when the Idle Time limit has been reached:
[PWXPC_10072] [INFO] [CDCDispatcher] session ended after waiting for [idle_time]
seconds. Idle Time limit is reached

This message is also issued when a continuous extraction is stopped with the PowerExchange STOPTASK
command. In this case, the idle_time variable in the message reports 86400, which is the “never expire” time
limit that is used when an Idle Time value of -1 is specified.

Tip: On very active systems, a positive value for Idle Time might never match. Use 0 if you do not want the
session to run continuously.

For example, if you specify an Idle Time value of 10 seconds and PowerExchange finds no data for the
sources in the change stream for a 10-second period, PowerExchange returns an EOF to the PowerCenter
Integration Service, which causes the session to terminate successfully.

If you specify values for Reader Time Limit and Idle Time, the PowerCenter Integration Service stops
reading data from the source when one of these terminating conditions is reached, whichever one is reached
first. If the Reader Time Limit is reached prior to the Idle Time limit, the session stops at that point, even
though the Idle Time limit has not yet been met.

Warning: Reader Time Limit does not result in normal termination of a CDC session. Use Idle Time instead
of Reader Time Limit.

120 Chapter 4: Connections

Configuring CDC Restart Attributes
The following table describes the optional restart connection attributes for CDC sessions:

Connection Change or Description

Attribute Real Time

Application Name Both Application name for the extraction. This name is part of the audit trail for
data extractions by the PowerExchange Listener.
The application name must be unique for each session. The name is case
sensitive and cannot exceed 20 characters.
Default is the first 20 characters of the workflow name.

RestartToken File Both Folder name that contains the restart token override file on the PowerCenter
Folder Integration Service node.
Default is $PMRootDir/Restart.

RestartToken File Both File name in the restart token file folder that contains the restart token
Name override file. PWXPC uses the contents of this file, if any, in conjunction with
the state information to determine the restart point for the session.
Default is the Application Name value if specified, or the workflow name if
Application Name is not specified.

You can use numerous CDC reader application connection attributes to specify restart information. PWXPC
uses the restart information to tell PowerExchange the point from which to start reading the captured change
data.

Warning: Be careful when you use the default value for Application Name. The default might not result in a
unique application name. The Application Name value and the RestartToken File Name values must be
unique for every CDC session. If a non-unique name is specified for either attribute, CDC session failures
and potential data loss can occur or results are unpredictable.

Related Topics:
• “Configuring the Restart Token File” on page 173

Configuring UOW Count

The following table describes the optional UOW Count connection attribute:

Connection Change or Description

Attribute Real Time

UOW Count Both The number of units of work (UOWs) that PWXPC processes before it sends
a commit request to PowerCenter to commit data to all targets in the CDC
session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before sending a
commit request.
Default is 1.

A unit of work (UOW) is a collection of changes within a single commit scope that a transaction makes on the
source system. Each UOW can contain a different number of changes.

Configuring Connection Attributes 121

 PWXPC reads change data from PowerExchange and provides the data to the source qualifier, at which point
the count of UOWs begins. When you use a non-zero value for the UOW Count attribute, PWXPC issues a
real-time flush to commit the change data to the target after the UOW count reaches the specified value.
When PWXPC issues a real-time flush, it writes the following message to the session log:
[PWXPC_10081] [INFO] [CDCDispatcher] raising real-time flush with restart tokens
[restart1_token], [restart2_token] because UOW Count [uow_count] is reached

For example, if the UOW Count value is 10, the PowerCenter Integration Service commits all change data
read from the source to the target after the tenth UOW is processed.

The lower you set the UOW Count value, the faster the PowerCenter Integration Service commits data to the
target. If you require the lowest possible latency for applying change change data to targets, specify a UOW
Count value of 1.

Commit processing is not controlled solely by the UOW Count attribute. The Maximum Rows Per commit,
and Real-Time Flush Latency attributes also determine the commit frequency.

If you have many small UOWs, you can use the UOW Count or Minimum Rows Per commit attribute or
both to create a consistent flow of UOWs that can be committed to the target in batches of approximately the
same size.

You can also control commits that occur at UOW boundaries based on time by specifying the Real-time
Flush Latency attribute. Specify the UOW Count or Real-time Flush Latency attribute or both. If you
specify both, a commit occurs when either limit is reached.

Warning: If you specify a low UOW Count value, the session might consume more system resources on the
target platform because data is committed to the target more frequently. Balance performance and resource
consumption with latency requirements when setting the UOW Count, Maximum Rows Per commit, and
Real-Time Flush Latency connection attributes.

Related Topics:
• “Commit Processing for CDC Sessions” on page 161

Configuring PWX Latency and Real-Time Flush Latency

The following table describes the optional flush latency connection attributes:

Connection Change or Description

Attribute Real Time

Real-Time Flush Real Time The number of milliseconds that must elapse before PWXPC sends a
Latency in milli- commit request to PowerCenter. When this period elapses, PWXPC
seconds continues to read the changes in the current UOW until it reaches the end of
the UOW. Then PWXPC sends a commit request to PowerCenter so that the
data can be committed to the targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse before
PWXPC sends a commit request to PowerCenter. If you specify a value from
0 through 2000, PWXPC uses 2000 milliseconds (2 seconds).
Default is 0.

PWX Latency in Real Time The maximum number of seconds that PowerExchange on the source
seconds machine waits for more change data before it flushes data to PWXPC on the
PowerCenter Integration Service machine.
Default is 2.

122 Chapter 4: Connections

When a CDC session begins, PWXPC passes the PWX Latency in seconds attribute value from the
connection to the PowerExchange Listener and requests change data from the change stream. The
PowerExchange Listener uses this latency value as the Consumer API (CAPI) interface timeout value on the
source machine. PowerExchange issues the following message on the source machine to identify the CAPI
interface timeout value:
PWX-09957 CAPI i/f: Read times out after <number> seconds

If you select the Retrieve PWX Log Entries attribute on the connection, PWXPC also writes this
informational message to the session log.

PowerExchange reads change data into a buffer on the source machine. When the one of the following
conditions occur, PowerExchange flushes the buffer that contains the change data to PWXPC on the
PowerCenter Integration Service machine:

• The buffer becomes full.

• A commit point occurs.
• The PWX Latency in seconds interval, or CAPI interface timeout, expires.

Note: The CAPI interface timeout also affects the speed with which PWXPC processes stop requests from
PowerCenter. PWXPC must wait for PowerExchange to return control to it before it can handle a stop
request.

After PowerExchange flushes change data to PWXPC, PWXPC provides the change data to the source
qualifier and the Real-Time Flush Latency interval begins. After the Real-Time Flush Latency interval
expires and PWXPC reaches a UOW boundary, PWXPC sends a commit request to PowerCenter and writes
the following message to the session log:
[PWXPC_10082] [INFO] [CDCDispatcher] raising real-time flush with restart tokens
[restart1_token], [restart2_token] because Real-time Flush Latency [milliseconds]
occurred

For example, if the value for real-time flush latency interval is 10 seconds, PWXPC sends a commit request
to PowerCenter for all complete UOWs after 10 seconds elapses and the next UOW boundary is met.

The lower you set the Real-Time Flush Latency interval, the faster change data is committed to the targets.
If you require the lowest possible latency for applying changes to targets, specify a low value for the Real-
Time Flush Latency interval.

Warning: When you specify low flush latency intervals, the CDC session might consume more system
resources on the source, PowerCenter Integration Service, and target systems for the following reasons:

• PowerExchange returns more frequently to PWXPC. As a result, PowerExchange passes fewer rows to
PWXPC on each iteration and consumes more resources on the source machine.
• PowerCenter commits change data to the targets more frequently. As a result, more resources are
consumed on the PowerCenter Integration Service and target machines.

When you set the flush latency intervals, balance performance and resource consumption with latency
requirements.

To control commit processing on the target, you can use the Real-Time Flush Latency attribute in
conjunction with the Maximum Rows Per commit and UOW Count connection attributes.

Configuring Connection Attributes 123

 Related Topics:
• “Commit Processing for CDC Sessions” on page 161

Configuring Maximum Rows per Commit

The following table describes the optional Maximum Rows Per commit connection attribute:

Connection Change or Description

Attribute Real Time

Maximum Rows Both The maximum number of change records that PWXPC processes before
Per commit it sends a commit request to PowerCenter to commit data to all targets in
the CDC session. PWXPC does not wait for a UOW boundary to commit
the change data.
Default is 0, which disables this attribute.

You can use the Maximum Rows Per commit attribute to commit change records between UOW
boundaries. Use this attribute when you have extremely large UOWs in the change stream that might cause
locking issues on the target database or high resource consumption on the PowerCenter Integration Service
machine. PWXPC uses the Maximum Rows Per commit value to commit data to the targets before the end
of a UOW is received, a process called a subpacket commit.

Important: Because PWXPC can commit change data to targets between UOW boundaries, referential
integrity (RI) might be compromised. Do not use this connection attribute if you have targets in the CDC
session with RI constraints.

The Maximum Rows Per commit value is the number of change records in a source UOW that PWXPC
processes before issuing a real-time flush to commit the change data to the target. When PWXPC flushes
change data to the target, it writes the following message to the session log:
[PWXPC_12128] [INFO] [CDCDispatcher] raising real-time flush with restart tokens
[restart1_token], [restart2_token] because Maximum Rows Per commit [commit_count]
occurred

For example, if each UOW contains 1,000 change records and you specify 100 for Maximum Rows Per
commit, PWXPC issues a real-time flush to commit the change records after each 100 records, which results
in 10 commits for each UOW.

As part of the commit processing, any locks in the target databases for these changes are released.

PWXPC resets the Maximum Rows Per commit counter when either the specified number of records is met
or the end of the UOW is reached.

Commit processing is not controlled solely by the Maximum Rows Per commit attribute. The UOW Count
and Real-Time Flush Latency attributes also determine the target commit frequency. The Maximum Rows
Per commit attribute differs from the UOW Count attribute in that it specifies a count of records within a
UOW whereas the UOW Count specifies a count of complete UOWs.

By using a subpacket commit for large UOWs, you can minimize lock contention on the target and reduce
storage use on the PowerCenter Integration Service machine. However, if you specify a low Maximum Rows
Per commit value, the session might consume more system resources on the target machine because
PWXPC sends commit requests to the target more frequently. Balance performance and resource
consumption with latency requirements when setting theMaximum Rows Per commit, UOW Count, and
Real-Time Flush Latency values.

When a session processes changes for multiple sources, the Maximum Rows Per commit count is
cumulative across all sources in the group. PWXPC sends a commit request to PowerCenter when the
maximum count of change records is reached, regardless of the number of sources to which the changes

124 Chapter 4: Connections

apply. For example, assume that a UOW contains 900 change records for one source followed by 100
change records for a second source and then another 500 change records for the first source. If you set the
Maximum Rows Per commit attribute to 1000, PWXPC issues the commit after the thousandth change
record, that is, after the 100 changes for the second source.

Warning: If a UOW contains changes for multiple source tables, the Maximum Rows Per commit attribute
might cause commits to be generated at points in the change stream where the relationship linking the
source tables is inconsistent. In this situation, target commit failures can occur.

Related Topics:
• “Commit Processing for CDC Sessions” on page 161

Configuring Minimum Rows per Commit

The following table describes the optional Minimum Rows Per commit connection attribute:

Connection Change or Description

Attribute Real Time

Minimum Rows Real Time The minimum number of change records that the PowerExchange
Per commit Listener must read from the change stream before it passes a commit
record to PWXPC. Before reaching this minimum, the PowerExchange
Listener passes only change records to PWXPC and discards any
commit records.
Default is 0, which disables this attribute.

If the change stream contains many small UOWs, you can use the Minimum Rows Per commit attribute to
create larger UOWs of a more uniform size.

This attribute specifies the minimum number of change records that the PowerExchange Listener must read
from the change stream before it passes a commit request to PWXPC. Until the minimum rows limit is met,
the PowerExchange Listener discards any commit records that it reads from the change stream and passes
only change records to PWXPC. After the minimum rows limit is met, PowerExchange passes the next
commit record to PWXPC and then resets the minimum rows counter.

Online transactions that run in transaction control systems such as CICS and IMS often issue a commit after
making only a few changes, which results in many, small UOWs in the change stream. PowerExchange and
PWXPC can process fewer, larger UOWs more efficiently. Therefore, by using this attribute to increase the
size of UOWs, you can improve CDC processing efficiency.

A minimum rows limit does not damage the referential integrity of the change data because PowerExchange
does not create new commit points in the change stream. PowerExchange only skips some of the original
commit records in the change stream.

Configuring Connection Attributes 125

 Connection Attributes for CDC Commit Processing
The following table describes the optional CDC connection attributes that you can use to control target
commit processing:

Connection Real Description

Attribute Time or
Change

Maximum Rows Both The maximum number of change records that PWXPC processes before it
Per commit sends a commit request to PowerCenter to commit data to all targets in the
CDC session. PWXPC does not wait for a UOW boundary to commit the
change data.
Default is 0, which disables this attribute.

Minimum Rows Real Time The minimum number of change records that PowerExchange must read from
Per commit the change stream before it can pass any source commit record to PWXPC.
Before reaching this minimum, PowerExchange passes only the change
records, without any commit records, to PWXPC.
Default is 0, which disables this attribute.

Real-time Flush Real Time The number of milliseconds that must elapse before PWXPC sends a commit
Latency in milli- request to PowerCenter. When this period elapses, PWXPC continues to read
seconds the changes in the current UOW until it reaches the end of the UOW. Then
PWXPC sends a commit request to PowerCenter so that the data can be
committed to the targets.
Valid values:
-1. Disables this attribute.
0 to 86400. Specifies the number of milliseconds that must elapse before
PWXPC sends a commit request to PowerCenter. If you specify a value from 0
through 2000, PWXPC uses 2000 milliseconds (2 seconds).
Default is 0.

UOW Count Both The number of UOWs that PWXPC processes before it sends a commit
request to PowerCenter to commit data to all targets in the CDC session.
Valid values:
-1. Disables this attribute.
0. Disables this attribute.
n. Specifies the number of UOWs that PWXPC processes before sending a
commit request.
Default is 1.

You can specify values for all of these commit control attributes. However, PWXPC flushes the data buffer to
commit change data to the targets only when one of the following attribute values is met, whichever is first:

• Maximum Rows Per commit

• Real-time Flush Latency
• UOW Count
After PWXPC commits the change data, it resets the Maximum Rows Per commit, Real-time Flush
Latency, and UOW Count counters and continues to read change data. When one of these commit-control
attribute values is met again, PWXPC commits the new change data to the targets. Commit processing
continues until the CDC session is stopped, ends normally, or ends abnormally. When the PWXPC CDC
reader ends normally, PWXPC issues a final commit to flush all of the complete, buffered UOWs and their

126 Chapter 4: Connections

 final restart tokens to the targets. Before ending, the PWXPC CDC reader writes the following message to the
session log:
PWXPC_12075 [INFO] [CDCRestart] Session complete. Next session will restart at:
Restart 1 [restart1_token] : Restart 2 [restart2_token]

Restriction: If you select the Commit On End Of File property on the session Properties tab, duplicate data
might occur on the targets because the PowerCenter Integration Service commits any remaining change data
in the buffer to the targets. This final commit occurs after the PWXPC CDC reader has committed all
complete UOWs in the buffer, along with their restart tokens, to the targets. As a result, the restart tokens
might represent a point in the change stream that is earlier than final change data that the PowerCenter
Integration Service commits to the targets. To prevent possible duplicate data when you restart CDC
sessions, set the Commit Type session property to Source and clear the Commit On End Of File property.

Related Topics:
• “Commit Processing for CDC Sessions” on page 161

Target Latency
Target latency is the total time that PWXPC takes to extract change data from the change stream and that
the PowerCenter Integration Service takes to apply that data to the targets. If this processing occurs quickly,
target latency is low.

The values that you use for the commit control attributes affect target latency. You must balance target
latency requirements with resource consumption on the PowerCenter Integration Service machine and the
target databases.

Lower target latency results in higher resource consumption because the PowerCenter Integration Service
must flush the change data more frequently and the target databases must process more commit requests.

The following default values for commit control attributes result in the lowest latency:

• 0 for Maximum Rows Per commit, which disables this option

• 0 for Minimum Rows Per commit, which disables this option
• 0 for Real-time Flush Latency in milli-seconds, which is equivalent to 2000 milliseconds or 2 seconds
• 1 for UOW Count
These values can decrease target latency because PWXPC commits changes after each UOW or on UOW
boundaries. However, these values also cause the highest resource consumption on the source system, the
Integration Service machine, and the target databases. Or these values might decrease throughput because
change data flushes too frequently for the PowerCenter Integration Service or the target databases to handle.

To lower resource consumption and potentially increase throughput for CDC sessions, specify a value
greater than the default value for only one of the following attributes:

• Minimum Rows Per commit

• Real-time Flush Latency in milli-seconds
• UOW Count
Disable the unused attributes.

Examples of Using CDC Commit Processing Attributes

The following examples show how the commit control attributes affect CDC commit processing with PWXPC.

Configuring Connection Attributes 127

 Example 1. Subpacket Commit and UOW Count
This example uses the Maximum Rows Per commit and UOW Count attributes to control target commit
processing. The change data consists of UOWs of the same size. Each UOW contains 1,000 change records.
The commit control attributes have the following values:

• Maximum Rows Per commit is set to 300.

• Minimum Rows Per commit is set to 0, which disables this attribute.
• Real-time Flush Latency in milli-seconds is set to 0, which causes 2 seconds to be used.
• UOW Count is set to 1.
Based on the maximum rows value, PWXPC flushes the data buffer after reading the first 300 records in a
UOW. This action commits the change data to the targets. PWXPC continues to commit change data to the
targets every 300 records.

PWXPC commits change data on UOW boundaries only for the UOW count and real-time flush latency
interval. If the real-time flush latency interval expires before PWXPC reads 300 change records, PWXPC still
commits change data based on the maximum rows value because that threshold is met before a UOW
boundary occurs.

When the end of the UOW is read, PWXPC commits the change data because the UOW Count value is 1.
PWXPC resets the UOW and maximum rows counters and the real-time flush latency timer each time
PWXPC issues a commit. Because all of the UOWs have the same number of change records, PWXPC
continues to read change data and commit change data to the targets at the same points in each UOW.

In this example, PWXPC commits change data at the following points:

• 300 change records based on the maximum rows value

• 600 change records based on the maximum rows value
• 900 change records based on the maximum rows value
• 1,000 change records based on the UOW count value

Example 2. UOW Count and Time-Based Commits

This example uses the UOW Count and Real-time Flush Latency in milli-seconds attributes to control
commit processing. The change data consists of UOWs of varying sizes. The commit control attributes have
the following values:

• Maximum Rows Per commit is set to 0, which disables this attribute.

• Minimum Rows Per commit is set to 0, which disables this attribute.
• Real-time Flush Latency in milli-seconds is set to 5000, which is equivalent to 5 seconds.
• UOW Count is set to 1000.
Initially, PWXPC reads 900 complete UOWs in 5 seconds. At that point, the real-time flush latency interval
has expired. As a result, PWXPC flushes the data buffer to commit change data to the targets. PWXPC then
resets both the UOW counter and real-time flush latency timer. When PWXPC reaches the thousandth UOW,
PWXPC does not commit change data to the targets because the UOW counter was reset to 0 after the last
commit.

PWXPC reads the next 1,000 UOWs in 4 seconds, which is less than the real-time flush latency timer.
PWXPC commits this change data to the target because the UOW counter has been met. PWXPC then
resets the real-time flush latency timer and the UOW counter.

PWXPC continues to read change data and commit change data to the targets based on the UOW count or
real-time flush latency limit, whichever is met first.

128 Chapter 4: Connections

 In this example, PWXPC commits change data at the following points:

• After UOW 900 because the real-time latency flush latency timer was met first.
• After UOW 1,900 because the UOW count was met first during the second commit cycle.

Example 3. Minimum Rows and UOW Count

This example uses the Minimum Rows Per commit and UOW Count attributes to control target commit
processing. The change data consists of UOWs of the same size. Each UOW contains ten change records.
The commit control attributes have the following values:

• Maximum Rows Per commit is set to 0, which disables this attribute.

• Minimum Rows Per commit is set to 100.
• Real-time Flush Latency in milli-seconds is set to -1, which is disables this attribute.
• UOW Count is set to 10.
PWXPC passes the minimum number of rows value to PowerExchange and requests change data from the
change stream. Because the minimum rows value is 100, PowerExchange skips commit records in the first 9
UOWs. When PowerExchange reads the last change record in the tenth UOW, the minimum rows limit is met.
PowerExchange then passes the commit record for the tenth UOW to PWXPC and resets the minimum rows
counter. PWXPC increases the UOW counter to 1.

PowerExchange and PWXPC continue to read the change data until the UOW counter is 10. At this point,
PWXPC flushes the data buffer to commit change data to the targets and resets the UOW counter.

In this example, PWXPC commits change data after 1,000 change records. This frequency is equivalent to
after every 10 UOWs because each UOW contains 100 change records and the UOW Count is 10.

Connection Attributes for Multiple-Record Writes

When you create a connection for a source or target in a workflow that performs multiple-record writes,
observe the following rules and guidelines:

• You can use multiple-record writes with PWXPC connections only. Multiple-record writes with
PowerExchange ODBC connections are not supported.
• Select Filter After for the Offload Processing source and target connection attribute. PowerExchange
and PWXPC perform offload processing on the Integration Service machine.
If you select a different value for Offload Processing, PowerExchange changes the value to Filter After.
However, if RBAs are being read using the GetDatabaseKey or GetIMSRBAByLevel function from an IMS
unload file in the BMC LONG format, you must select No for the Offload Processing source connection
attributes. Otherwise, the session will fail with message PWX-03803.
• Multithreaded processing is supported for IMS source connections if you select Filter After for the
Offload Processing connection attribute and set the Worker Threads connection attribute to a non-zero
value.
• Multithreaded mapping processing is not supported for IMS target connections. If you set the Worker
Threads connection attribute to a non-zero value, the session ignores this setting.
• Multithreaded processing is not supported for sequential or VSAM connections.
• Multithreaded processing is not supported for sequential or VSAM source or target connections.
If you set the Worker Threads connection attribute to a nonzero value, the setting is ignored.

Connection Attributes for Multiple-Record Writes 129

 • Select Off for the Confirm Write source and target connection properties. If you select a different value,
PowerExchange changes the value to Off.
• In certain cases, you might need to change the value of the CSQ_MEMSIZE parameter of the PWX
Override connection property. This property defines the maximum amount of memory that the cache can
consume for multiple-record writes.

Related Topics:
• “Multiple-Record Writes to Nonrelational Targets” on page 53
• “Considerations for Sessions that Perform Multiple-Record Writes” on page 135

Asynchronous Network Communication

PowerExchange uses asynchronous network communication for all send and receive data operations
between a PowerExchange client and a PowerExchange Listener. With asynchronous communication,
PowerExchange uses separate threads for network processing and data processing, so that network
processing overlaps with data processing.

To detect unsuccessful connection attempts or failed connections, PowerExchange provides several types of
timeout processing, including heartbeat timeout processing. During periods of inactivity, the network threads
on the PowerExchange client and PowerExchange Listener send and receive heartbeat data. If no heartbeat
or non-heartbeat data is sent or received during the heartbeat timeout interval, PowerExchange terminates
the connection and issues an error message.

PowerExchange asynchronous communication, including heartbeat processing, provides the following

benefits:

• Improved session performance for streaming data transfers

• Early detection of connection failure situations
• Timeout processing that does not require configuration in most cases

Asynchronous Read and Write Processing

Asynchronous read and write operations use application threads, network threads, and circular buffers to
exchange data between a PowerExchange client and PowerExchange Listener.

The following steps summarize the asynchronous read process:

1. The application thread on the PowerExchange Listener writes data to the circular buffer on the
PowerExchange Listener.
2. The network thread on the PowerExchange Listener reads data from the circular buffer and sends the
data across the network.
3. The network thread on the PowerExchange client reads data from the network and writes the data to the
circular buffer on the PowerExchange client.
4. The application thread on the PowerExchange client reads the data from the circular buffer and
processes the data.
The following steps summarize the asynchronous write process:

1. The application thread on the PowerExchange client writes data to the circular buffer on the
PowerExchange client.

130 Chapter 4: Connections

 2. The network thread on the PowerExchange client reads data from the circular buffer and sends the data
across the network.
3. The network thread on the PowerExchange Listener reads data from the network and writes the data to
the circular buffer on the PowerExchange Listener.
4. The application thread on the PowerExchange Listener reads the data from the circular buffer and writes
the data to the target.

Timeout Processing
To detect unsuccessful connection attempts or failed connections, PowerExchange provides the following
types of timeouts:

Connection timeouts

Connection timeouts are used to detect an unsuccessful connection attempt. By default, PowerExchange
tries to connect for 180 seconds. If a connection cannot be established, PowerExchange issues an error
message. You can change the connection timeout interval by using the PWX Override connection
attribute.

Heartbeat timeouts

During periods of inactivity, the network threads on the PowerCenter client and PowerCenter Listener
send and receive heartbeat data. If no data, including heartbeat data, is sent or received over an interval
of 500 seconds, PowerExchange terminates the connection and issues an error message.

Network operation timeouts

You can optionally configure network operation timeouts. When the client-side application thread
perceives a single network operation to exceed the specified number of seconds, PowerExchange ends
the connection and issues a timeout error message.

By default, network operation timeouts are not enabled, and PowerExchange uses heartbeat processing
only to detect send and receive timeouts.

Note: The TIMEOUTS statement in the DBMOVER configuration file is deprecated and no longer has any
effect.

Configuring Connection and Network Operation Timeouts

Use this procedure to override the default connection timeout or to enable a network operations timeout.

To configure connection and network operation timeouts:

1. In the Workflow Manager, connect to a PowerCenter repository.

2. Complete one of the following actions:
• Click Connections > Relational.
• Click Connections > Application.
The Relational Connection Browser dialog box or Application Connection Browser dialog box
appears.
3. In the Select Type field, select the type of connection for which you want to configure timeouts.
4. Click Edit or New, based on whether you are creating or editing a connection.
The Connection Object Definition dialog box appears.
5. Scroll to the PWX Override connection attribute. In the Value column, specify the number of seconds for
the connection timeout interval, network operation timeout interval, or both.

Asynchronous Network Communication 131

 If you specify both overrides, in either order, separate them with a semicolon. Use the following syntax:
TCPIP_OP_TIMEOUT=network_operation_timeout;TCPIP_CON_TIMEOUT=connection_timeout
6. Click OK.

132 Chapter 4: Connections

CHAPTER 5

Working with Sessions

This chapter includes the following topics:

• Working with Sessions Overview, 133

• Extracting Data in Batch Mode, 133
• Extracting CDC Data in Change and Real-time Modes, 146
• Loading Data to PowerExchange Targets, 153
• Constraint-Based Loading for Relational Targets, 157
• Variables for Session and Connection Properties, 158
• Workflows that Connect to a PowerExchange Listener Service, 159
• Configuring PowerExchange Services for Workflows, 160
• Pipeline Partitioning, 160
• Commit Processing for Bulk Data Movement Sessions, 160
• Commit Processing for CDC Sessions, 161
• Considerations for Pre- and Post-Session SQL Commands, 161
• Preserving Low Values in Source Character Fields, 162

Working with Sessions Overview

After you create mappings in the PowerCenter Designer, you can create a session and use the session in a
workflow to extract, transform, and load data. You create sessions and workflows in the Workflow Manager.

You can create a session in a workflow to extract data in batch, change, or real-time mode. You determine
how you want the PowerCenter Integration Service to extract the data when you configure the session. You
can also create a session to load data to a target.

Extracting Data in Batch Mode

To extract data in batch mode, you must select a connection for batch mode and configure session
properties. The connection you select and the session properties you configure differ depending on the type
of source data.

133
 Data Extraction from Multiple Sequential or VSAM ESDS Files
You can extract data from multiple sequential or VSAM ESDS files with the same record layout by using
PowerExchange file list processing.

You must use a filelist file. A filelist file specifies the names of the files from which source data is to be
extracted. The PowerExchange Listener extracts data from all of the files named in the filelist file in the order
in which the files are listed.

To request file list processing for a sequential or VSAM ESDS source, configure the following session
attributes:

• Specify the filelist file name in the File Name Override attribute.
• Select the Filelist File option.

Related Topics:
• “Configuring Batch Mode Sessions for Nonrelational Sources” on page 138

Overrides for the IMS Access Method and Related Properties

You can override the access method for a PowerExchange IMS source or target in the session properties.
You can also override related session properties, such as IMS SSID, PSB name, PCB name, and PCB
number depending on the access type.

Use overrides for the following reasons:

• Gain greater flexibility in configuring bulk data movement sessions. When you are ready to run a session,
you can switch between the IMS ODBA and DL/1 BATCH access methods as needed. Use ODBA to run
small, frequent database queries as PowerExchange Listener subtasks to read or write data. Run DL/I
batch or BMP jobs at night to process mass updates that require longer access to the database.
• Create fewer PowerExchange data maps. You do not need to create a duplicate data map to change the
access method or a related property. With fewer data maps, you also can create fewer PowerCenter
mappings and workflows.
• Avoid creating very large PSBs that include many PCBs and require large buffers. Large PSBs are
sometimes used to avoid exceeding the maximum of ten NETPORT statements in the DBMOVER
member. Instead, you can override the PSB and PCB number at session run time.
• Use one pair of NETPORT and LISTENER statements for DL/I or BMP access to multiple PSBs.
• Override the IMS SSID in an ODBA data map to point to a source or target on another IMS subsystem.
You can also use an IMS SSID override to direct a BMP netport job to another IMS SSID or to use another
DBD library for an IMS source unload file.
In PowerCenter, enter the session overrides on the Mapping tab of the Edit Tasks dialog box. The following
overrides are available for the source and target, depending on the access method:

• IMS AM Override. Overrides the access method.

• IMS SSID Override. Overrides the IMS SSID.
• IMS PSBNAME Override. Overrides the PSB name. Available for IMS ODBA and DL/I batch access.
• IMS PCBNAME Override. Overrides the PCB name. Available for IMS ODBA access only.
• IMS PCBNUMBER Override. Overrides the PCB number. Available for DL/I batch access only.

Note: In the PowerExchange Navigator, you can override these properties for a database row test of an IMS
data map.

134 Chapter 5: Working with Sessions

Considerations for Sessions that Perform Multiple-Record Writes
When you work with sessions that perform multiple-record writes, special considerations apply in the
following areas:

• Sequencing and queuing cache for multiple-record writes

• Row statistics for multiple-record writes
• Multiple-record write session properties for sources
• Multiple-record write session properties for targets

Related Topics:
• “Multiple-Record Writes to Nonrelational Targets” on page 53
• “Connection Attributes for Multiple-Record Writes” on page 129

Sequencing and Queuing Cache for Multiple-Record Writes

For multiple-record write operations, PowerExchange maintains a cache, called the sequencing and queuing
cache, on the PowerCenter Integration Service machine. PowerExchange uses this cache to queue data from
nonrelational sources with multiple record types in the correct sequence before sending it to the target.

A PowerCenter workflow processes data from different source record types in different pipelines. The source
records can represent IMS segments or sequential or VSAM records. The PowerExchange Writer receives
data for a given record type in order, but data for different record types might arrive out of order.

Data for any given record type arrives and is written to the cache the same sequence as on the source. The
relationship of data for any record type to other records is maintained by the following items:

• Primary key (PK). Contains a sequence number that indicates the physical sequence of the record in the
source.
• Foreign Key (FK), the sequence number of the data record of which the current data is a child
• segment sequence number, which corresponds to the segment or record to which the data belongs
You can use the CSQ_MEMSIZE parameter of the PWX Override connection parameter to limit the memory
that the cache uses. When this limit is reached, data is spilled to files on disk.

Separate threads write to and read from the cache.

The producer thread writes data records to the cache. The producer maintains one or more queues for each
record.

The consumer thread reads data from the cache, uses the sequencing information to reassemble the data in
the correct order, and writes the data to the target data set. The consumer thread reads records from the
cache in primary key sequence. Reading from the cache correctly reflects the original order in which the data
was read from the source.

If the consumer thread encounters any orphan records, it skips them and reports the number of records
skipped. In this way, the consumer thread detects situations where data has been orphaned by a parent that
was filtered out or dropped through the mapping logic.

PowerExchange reports the number of records that it writes to and reads from the cache in the following
messages:
PWX-35805 Records written to cache: total_record_count
PWX-35806 Records read from cache: total_record_count

Extracting Data in Batch Mode 135

 If PowerExchange skips any records, it reports the number of records that were cached, written, and skipped
in the following message:
PWX-35807 Records cached = cached, records written = written, records skipped = skipped

PowerCenter also provides messages that involve the cache. For more information, see “Row Statistics for
Multiple-Record Writes” on page 136.

Notes:

• Because of its asynchronous nature, the PowerExchange writer might still be writing data after
PowerCenter has processed its last commit.
• If a VSAM or sequential data set includes one or more header records, PowerExchange might write the
entire file into the cache and write to the target only when all records have been processed.
• Because data for different record types are processed by different PowerCenter pipelines, the amount of
data written into the cache largely depends on the arrival of data for each record type in PowerExchange.
For example, a record type that does not have any data at all might force the entire file to be written into
the cache before being processed.

Row Statistics for Multiple-Record Writes

The following statistics for multiple-record writes are reported in session logs and the Workflow Monitor:

• Requested row count. The number of rows that PWXPC has passed to the sequencing and queuing
cache.
• Applied and affected row count. The number of rows that have been successfully read from the
sequencing and queuing cache.
• Rejected row count. The number of rows that were rejected as orphans. Orphans are records for which no
record exists in the cache with a primary key that matches the foreign key of the orphan.
Because multiple-record writes require the Confirm Write Off target connection setting, these writes are
asynchronous with respect to PowerCenter. Depending on input data to the writer, row statistics reported in
the session log and Workflow Monitor might not always reflect the number of rows that have actually been
written to the target.

Multiple-Record Write Session Properties for Sources

For PowerCenter bulk data movement sessions that perform multiple-record writes, you can define the
following session properties for the source:

Flush After N Blocks

Optional. By default PWXPC flushes blocks of data only when buffers are completely full or at end-of-file.
Flushing might not occur frequently if some of the record types do not have as much data as others. A
delayed flush can cause one of the seldom-occurring record types to have no data to be written to the
target for a long time, thereby blocking flushing on the writer side.

To ensure that buffers for all records are flushed at a regular interval, define this session property. This
property specifies the maximum number of blocks that can be flushed across all record types without a
block for any one record type being flushed. A value of zero disables this feature. Flushing then occurs
only when blocks are filled.

Valid values are -1 to 100000.

The default value of -1 works in the following manner:

• For multiple-record sources that do not use sequencing, the default value of Flush After N blocks = 0.
That is, the feature is disabled by default.

136 Chapter 5: Working with Sessions

 • For multiple-record sources that use sequence fields, the default value of Flush After N blocks =
7*(number of record types in the source).

Example

Assume a multiple-record source has four record types: A, B, C, and D. If Flush After N Blocks is not
defined, the following blocks are flushed when their buffers are full:

C, C, C, B, C, B, C, C, C, C, C, B, A, D flush (all)

If Flush After N Blocks is set to 4, the following flushes occur:

C, C, C, B + ForcedFlush(A, D), C, B, C, C + ForcedFlush(A, D), C, C + ForcedFlush (B), C, A +

ForcedFlush (D), Flush (all)

The terminology ForcedFlush(X) indicates that the flush is caused by the Flush After N Blocks setting
and not by a block becoming full.

IMS Unload File Name

An IMS unload file name. Specify this attribute to read source data from an IMS unload file instead of
from the IMS database. Required when the Use Sequence Fields option is selected for an IMS source
definition.

File Name Override

Source file name. Optionally define this attribute for sequential or VSAM sources. This source file name
takes precedence over the file name in the source definition or data map.

Multiple-Record Write Session Properties for Targets

When you define a PowerCenter session that performs a multiple-record write, you must specify the name of
the target file in the target session properties. Optionally, enter the other session properties for the target.

File Name Override

Target sequential or VSAM file name. This file name overrides the file name in the target definition and
data map.

IMS Unload File Name

Target IMS unload file name. Specify this attribute to write source data to an IMS unload file instead of
an IMS database. Required when the Use Sequence Fields option is selected for an IMS target
definition.

Insert SQL override

Overrides the SQL query for inserts that is sent to PowerExchange.

PWXPC replaces the default SQL query with the SQL statement that you enter. PWXPC passes this
SQL statement to PowerExchange for processing. You can enter any SQL statement that
PowerExchange supports for NRDB SQL.

For a multiple-record target with use sequence fields, the SQL query override must include the
generated sequence fields in the query.

Use the following syntax:

group_name1=insert_sql_override1; group_name2=insert_sql_override2;...

Map Name Override

Overrides the name of the PowerExchange data map for the target.

PWXPC sends the map name from the source definition unless this value overrides it.

Extracting Data in Batch Mode 137

 Pre SQL
One or more SQL statements that PWXPC executes before the session runs with the target database
connection.

Post SQL
One or more SQL statements that PWXPC executes after the session runs with the target database
connection.

Schema Name Override

Overrides the schema name of the PowerExchange data map for the target.

PWXPC sends the schema name from the source definition unless this value overrides it.

SEQ Properties on z/OS

For SEQ targets on z/OS, you can optionally define the following data set allocation properties to have
the session allocate the target data set when the session runs:

• Disp
• Space
• Primary Space
• Secondary Space
• LRECL
• BLKSIZE
• RECFM
• VOLSER
• MGMTCLAS
• DATACLAS
• STORCLAS
• MODELDSCB

For more information about session properties for targets, see “Configuring Sessions to Load Data to
Nonrelational Targets” on page 153.

Configuring Batch Mode Sessions for Nonrelational Sources

When you configure a session to extract data from a nonrelational source in batch mode, you can override
data map information for the source. PWXPC and PowerExchange use this information to extract the source
data.

1. In the Task Developer, double-click a session with a nonrelational source.

2. On the Mapping tab, expand the Sources node in the Transformations tree. Then click a nonrelational
source.
Under Readers, PowerCenter sets the PowerExchange batch reader for the source type. For
nonrelational batch sources, the reader name has one of the following formats:
PowerExchange Batch Reader for database_type
PowerExchange Reader for database_type
The database_type variable represents one of the following values:
• ADABAS

138 Chapter 5: Working with Sessions

 • ADABAS Unload Files
• DB2 Datamaps
• DB2 Unload Datasets
• DATACOM
• IDMS
• IMS
• Sequential Files
• VSAM Files
Note: The name of the reader cannot be altered. The only exception to this rule is for Adabas. You can
use either ADABAS or ADABAS Unload Files.
3. Under Connections, in the Value field, select a PWX NRDB Batch connection.
4. Under Properties, set the session attributes that apply to the source type, as needed.
The following table describes these attributes:

Attribute Name Source Type Description

Schema Name Override All Overrides the schema name in the source
PowerExchange data map.

Map Name Override All Overrides the data map name of the source
PowerExchange data map.

File Name ADABAS Unload Specifies the file name of the unloaded Adabas
database.
Required for ADABAS Unload.

ADABAS Password ADABAS Password for the ADABAS database.

If the ADABAS FDT for the source file is password-
protected, enter the ADABAS FDT password.

Database Id Override ADABAS, ADABAS Overrides the ADABAS Database ID in the

Unload PowerExchange data map.

File Id Override ADABAS, ADABAS Overrides the Adabas file ID in the PowerExchange
Unload data map.

DB2 Sub System Id DB2 Datamaps Overrides the DB2 subsystem ID in the
PowerExchange data map.

DB2 Table name DB2 Datamaps Overrides the DB2 table name in the
PowerExchange data map.

Unload File Name DB2 Unload Datasets Overrides the DB2 unload file name in the
PowerExchange data map.

Extracting Data in Batch Mode 139

 Attribute Name Source Type Description

Filter Overrides All Filters the source data that PowerExchange reads
based on specific conditions that you define.
PWXPC adds the filter conditions in a WHERE
clause on a SELECT SQL statement and then
passes the SQL statement to PowerExchange for
processing. You can use any filter condition syntax
that PowerExchange supports for NRDB SQL.
For a single-record source, use the following syntax:
filter_condition
For example, the following filter condition selects
records where a column called TYPE has a value of
A or D:
TYPE=‘A’ or TYPE=‘D’
For a multiple-record source, use one of the
following syntax alternatives:
filter_condition
group_name1=filter;
group_name2=filter;...
The group_name syntax limits the SQL query
condition to a specific record in a multi-record
source definition. If you do not use the group_name
syntax, the SQL query condition applies to all
records in the multi-record source definition.
For example, to select only records that contain an
ID column value of "DBA" for a multi-record source
that has USER1 and USER2 records, specify one of
the following SQL query conditions:
USER1=ID=’DBA’;USER2=ID=’DBA’
ID=’DBA’

IMS Unload File Name IMS An IMS database unload file name. Required if you
want to read source data from the backup file
instead of from the IMS database. For a multiple-
record write to an IMS unload file, required for both
the source and target.

140 Chapter 5: Working with Sessions

Attribute Name Source Type Description

IMS AM Override IMS Overrides the IMS access method in the imported
data map for the source with the other available
access method. The session then uses the override
access method at run time.
- If you imported a source data map that specifies the
DL/1 BATCH access method, enter O to override it
with the IMS ODBA access method. For ODBA
access, you must also specify the IMS PSBNAME
Override and IMS PCBNAME Override attributes.
- If you imported a source data map that specifies the
IMS ODBA access method, enter D to override it with
the DL/1 BATCH access method, which provides
DL/I or BMP access. You must also specify the IMS
PCBNUMBER Override attribute.
Important: Before you run the session with an
access method override, ensure that you complete
the PowerExchange configuration tasks for the new
access method. For example, if the override is DL/1
BATCH, you must configure LISTENER and
NETPORT statements in the DBMOVER member
and configure the netport JCL. If the override is IMS
ODBA, you must perform other configuration tasks.
For more information, see "IMS Data Maps" in the
PowerExchange Navigator User Guide.

IMS SSID Override IMS If you imported an IMS ODBA data map for the
source and did not override the access method, use
this attribute to override the IMS subsystem ID
(SSID) from the data map for the session. If you
specified ODBA access as an override in the IMS
AM Override session attribute, you must enter this
value. An SSID is required for ODBA access.
If the session has an IMS unload file source, you
can use this override to point to another IMSID
statement in the DBMOVER member for the purpose
of changing from one DBD library to another DBD
library. By using the override, you can switch DBD
libraries without editing or adding any IMSID
statement and restarting the PowerExchange
Listener. For example, use this override to test
changes that you made to a DBD library against an
unload file.
If you use a netport job with BMP access to IMS,
you can use this override with the %IMSID
substitution variable in the netport JCL to specify an
IMS SSID to use for the session. This override
replaces the substitution variable. By using the
override with the substitution variable, you can use
the same netport JCL to access multiple IMS
environments, such as development, test, and
production environments.
Note: An IMS SSID is not required for DL/I batch
access to IMS data or for access to an IMS unload
file.

Extracting Data in Batch Mode 141

 Attribute Name Source Type Description

IMS PSBNAME IMS If you imported an IMS ODBA data map for the
Override source and did not override the access method, this
value overrides the PSB name from the data map. If
you specified ODBA access as an override in the
IMS AM Override attribute, you must enter this
value. A PSB name is required for ODBA access.
If you use DL/I batch or BMP access and specify
this override, you must also specify the
PSB=%PSBNAME substitution variable in the
netport JCL. The override value then replaces the
substitution variable in the JCL.
If you specify the PSB=%1 substitution variable
instead of PSB=%PSBNAME in the netport JCL, the
session uses the PSB name from the NETPORT
statement, if specified. In this case, you need a
separate NETPORT statement for each PSB. To
avoid exceeding the limit of ten NETPORT
statements in the DBMOVER member, use this
override with %PSBNAME substitution variable
instead.
Note: A PSB name is not used for access to an IMS
source unload file.

IMS PCBNAME IMS If you imported an IMS ODBA data map for the
Override source and did not override the access method, this
value overrides the PCB name from the data map. If
you specified ODBA access as an override in the
IMS AM Override attribute, you must enter this
value. A PCB name is required for ODBA access.
A PCB name is not used for DL/I batch or BMP
access or for access to an IMS unload file.

IMS PCBNUMBER IMS If you imported a DL/1 BATCH data map for the
Override source and did not override the access method, this
value overrides the PCB number from the data map.
If you specified DL/I access as an override in the
IMS AM Override attribute, you must enter this
value. A PCB number is required for DL/I or BMP
access.
A PCB number is not used for IMS ODBA access or
for access to an IMS unload file.

File Name Override VSAM, SEQ Overrides the data set or file name in the
PowerExchange data map.
Enter the complete data set or file name.
For the i5/OS, the format is: library_name/file_name.
If you select the Filelist File check box, enter the
name of a filelist file in this attribute. A filelist file is a
list of files.

Filelist File VSAM, SEQ Select this attribute only if you entered a filelist file
in the File Name Override field. A filelist file is a list
of files.

142 Chapter 5: Working with Sessions

Attribute Name Source Type Description

SQL Query Override All Overrides the SQL query sent to PowerExchange,
including any Filter Overrides.
PWXPC replaces the default SQL query with the
SQL statement that you enter and passes the SQL
statement to PowerExchange for processing. You
can enter any SQL statement that PowerExchange
supports for NRDB SQL.
For a single-record source, use the following syntax:
SQL_query_override
For example, you can select records from table
USER where a column called TYPE has a value of A
or D by specifying the following SQL query override:
Select ID, NAME from USER where
TYPE=‘A’ or TYPE=‘D’;
For a multiple-record source, use the following
syntax:
group_name1=sql_query_override1;
group_name2=sql_query_override2;...
For example, you can select only records with ID
column values that contain DBA for a multi-record
source with two records called USER1 and USER2
by specifying the following SQL query override:
USER1=Select ID, NAME from USER1
where ID='DBA'; USER2=Select ID,
NAME from USER2 where ID='DBA';

Extracting Data in Batch Mode 143

 Attribute Name Source Type Description

PWX Partition Strategy Offloaded DB2 Unload, Specifies one of the following partitioning strategies:
VSAM, and sequential - Single Connection. PowerExchange creates a
data sets single connection to the data source. Any overrides
specified for the first partition are used for all
partitions. With this option, if you specify any
overrides for other partitions that differ from the
overrides for the first partition, the session fails with
an error message.
- Overrides Driven. If the specified overrides are the
same for all partitions, PowerExchange creates a
single connection to the data source. If the overrides
are not identical for all partitions, PowerExchange
creates multiple connections.

Flush After N Blocks Multiple-record sources For bulk multiple-record sources, by default,
PWXPC flushes blocks of data only when the buffers
are completely full or at end-of-file. If some record
types do not have as much data as others, flushing
might not occur often. In this case, the record types
might not have data on the target for a long time,
thereby blocking flushes on the writer side.
To ensure that buffers for all record types are
flushed at a regular interval, define this Flush After
N Blocks session property. This property specifies
the maximum number of block flushes that can
occur across all record types without any one block
being flushed. A value of zero disables this feature
and causes flushing to occur only when blocks are
full.
Valid values for the property are -1 to 100000.
The default value of -1 works in the following
manner:
- For all multiple-record sources that do not use
sequence fields, process the same as Flush After N
Blocks = 0, which disables this feature and flushes
only when blocks are full .
- For all multiple-record sources that use sequence
fields, use Flush After N Blocks = 7 * (number of
record types in the source).

5. Click OK.

Related Topics:
• “Filtering Source Data with PWXPC” on page 210

Configuring Session Properties for Relational Batch Mode

Sessions
To extract data from a relational source in batch mode, select a PWXPC relational database connection. You
can then configure the session properties.

Session properties unique for PowerExchange are noted.

To configure session properties for a relational batch mode session:

1. In the Task Developer, double-click a session with a relational source to open the session properties.

144 Chapter 5: Working with Sessions

2. Click the Sources view on the Mapping tab.
3. In the Reader field of the Readers settings, select Relational Reader. For DB2 for z/OS, you can also
select a PowerExchange Reader for DB2 Image Copy reader.

4. If you use the relational reader, select one of the following relational database connections:
• For DB2 for i5/OS, select PWX DB2i5OS.
• For DB2 for Linux, UNIX, and Windows, select PWX DB2LUW for DB2.
• For DB2 for z/OS, select PWX DB2zOS.
• For Microsoft SQL Server, select PWX MSSQLServer.
• For Oracle, select PWX Oracle.
If you use the DB2 Image Copy reader, select a PWX NRDB Batch application connection.
5. Optionally, click the Sources view on the Mapping tab, and set attributes in the Properties area.
Caution: For DB2 for z/OS data sources, PowerExchange automatically appends FOR FETCH ONLY to
SQL SELECT statements. If you include FOR FETCH ONLY in the Sql Query attribute in the Properties
area, the expression is included twice in the SELECT statement, and PowerExchange issues an error.
The following table describes the attributes that you can configure for the PowerExchange Reader for
DB2 Image Copy reader:

Attribute Name Description

Schema Name Override Overrides the source schema name.

Map Name Override Overrides the source table name.

DB2 Sub System Id Overrides the DB2 instance name in the PowerExchange data map.

Image Copy Dataset Provides the image copy data set name. If not specified, the most current
image copy data set with TYPE=FULL and SHRLEVEL=REFERENECE is
used.

Disable Consistency If cleared, PowerExchange reads the catalog to verify that the DSN of the
Checking specified Image Copy Dataset is defined with SHRLEVEL=REFERENCE
and TYPE=FULL and is an image copy of the specified table. If the DSN is
not defined with these properties, the session fails.
If selected, PowerExchange reads the Image Copy Dataset regardless of
the values of SHRLEVEL and TYPE and without verifying that the object ID
in the image copy matches the object ID in the DB2 catalog.

Extracting Data in Batch Mode 145

 Attribute Name Description

Filter Overrides Filters the source data that PowerExchange reads based on specified
conditions.
PWXPC adds filter conditions specified to the WHERE clause on the
SELECT SQL statement and passes the SQL statement to
PowerExchange for processing. You can use any filter condition syntax
that PowerExchange supports for NRDB SQL. For more information, see
the PowerExchange Reference Manual.
For example, you can select records where a column called TYPE has a
value of A or D by specifying the following filter condition:
TYPE=‘A’ or TYPE=‘D’

SQL Query Override Overrides the SQL query sent to PowerExchange, including any filter
overrides.

6. Click OK.

Related Topics:
• “Filtering Source Data with PWXPC” on page 210

Extracting CDC Data in Change and Real-time

Modes
To extract change capture data in change mode (CAPX), you select a change mode application connection.
To extract data in real-time mode (CAPXRT), you select a real-time mode application connection.

Nonrelational source definitions imported in PowerCenter 8.x automatically have the appropriate Batch
Reader selected for the source type. This reader selection cannot be changed.

Related Topics:
• “CDC Application Connection Attributes” on page 116
• “Restart and Recovery Overview” on page 163
• “Commit Processing for CDC Sessions” on page 161
• “Understanding PWXPC Restart and Recovery” on page 164
• “Configuring the Restart Token File” on page 173
• “Configuring CDC Sessions” on page 178
• “Enabling Session Recovery for a CDC Session” on page 177

PowerExchange Extraction Maps

For CDC sessions, import source metadata from PowerExchange extraction maps to define source
definitions that can be used in mappings.

Extraction maps are generated by the PowerExchange Navigator when you create a capture registration for a
source. If necessary, you can modify an extraction map or create another extraction map based an existing
capture registration.

146 Chapter 5: Working with Sessions

 When you import an extraction map in PowerCenter Designer, select the CDC Datamaps option in the
Import from PowerExchange dialog box. Designer then creates the source definition based on the
extraction map. The Metadata Extensions tab for the source definition shows the schema name and map
name imported from the extraction map. You do not need to specify this information in the session properties.

Alternatively, you can import source metadata from a relational database to define source definitions.
However, you must then manually specify the extraction map name in the session Properties to use a CDC
reader and an application CDC connection.

Note: For nonrelational sources, you must import an extraction map to use a CDC reader. You cannot select
a CDC reader for sources that have a Database type value of PWX_source_NRDB2.

If you want to extract change data from a multi-record nonrelational source with extraction maps, you must
create a PowerExchange capture registration for every table in the data map. For each registration, the
PowerExchange Navigator creates a corresponding extraction map. You can then either import the data map
as a multi-record nonrelational source (for batch usage) or import the extraction map for each table (for CDC
usage).

Data Extraction from Multiple CDC Sources in a Session

If a mapping contains multiple sources from which you want to extract CDC data in change mode or real-time
mode, all sources must use the same application connection.

Specify the application connection only on the first source in the mapping. All of the other connections can
have a Type value of None.

Note: All of the sources must have the same source datatype. Sessions with mappings that include sources
with multiple datatypes fail at run time.

Session Properties for Change and Real-time Mode Sessions

When you configure a session to extract data in Change or Real-time mode, PWXPC uses the extraction map
to extract the source data. Source definitions for CDC sources can be created with extraction maps (CDC
Datamaps in the Import dialog box) or by importing metadata from the relational database. The session
properties will differ based upon how the source metadata was imported.

Configuring a Change or Real-time Session for Extraction Map Sources

Configure a Change or Real-time session for extraction map sources from the Task Developer.

1. In the Task Developer, double-click the session.

The Edit Tasks dialog box appears.
2. On the Properties tab, edit the following fields:
• In the Commit Type field, select Source.
• Clear the Commit on End of File property.
• To enable recovery for the session, set the Recovery Strategy property to Resume from last
checkpoint. This setting is important for CDC sessions because it ensures that data and restart
tokens are properly handled.
• If you write change data to a relational target, ensure that the Session Retry on Deadlock property
is not selected in the Performance section to prevent retries of target writes when a deadlock occurs.
By default, this property is not selected. Also, in the relational connection definition for the target,
ensure that the Connection Retry Period attribute is set to the default value of 0 to prevent

Extracting CDC Data in Change and Real-time Modes 147

 database connection retries. Changing either of these settings from the default values requires
enough buffers in the DTM buffer pool to hold the largest transaction that the session processes.
If you do not use the default settings for the Session Retry on Deadlock property and Connection
Retry Period attribute, the CDC session might hang with the following message because not enough
buffers are available:
WRT_8450 : The writer is waiting for buffer blocks. To improve session
performance and to prevent the session from hanging, increase the DTM buffer pool
size in the session properties.
In this case, set the Commit Interval session property and the Maximum Rows Per commit
attribute on the source connection to the same value, for example, 1000. Although the Commit
Interval property does not control when the session commits data to targets, it does affect the
number of buffers in the DTM buffer pool that can be used to hold a transaction. Use the Maximum
Rows Per commit connection attribute to divide the rows in a large transaction into subsets for
subpacket commit processing on the target. For more information about controlling commit
processing, see “Connection Attributes for CDC Commit Processing ” on page 126.
3. On the Mapping tab, click the Sources view.
4. In the right pane under Readers, for relational sources, select the appropriate CDC Reader.
With extraction map sources, PWXPC automatically selects the reader based on the source type of the
extraction map.
5. Under Connections, in the Value field, select CDC Real Time or CDC Change application connection
type.
PWXPC displays the valid connections for the source type in the Application Connection Browser.
Optionally, you can open the application connection to override any connection values.
6. Under Properties, you can configure optional attributes.
The following table describes the optional attributes that you can configure:

Attribute Source Description

Name Type

Schema Name All Overrides the schema name in the PowerExchange extraction map.
Override

Map Name All Overrides the PowerExchange extraction map name.

Override

ADABAS Adabas An Adabas password for the source file.

Password If the Adabas FDT for the source file is password-protected, enter the
Adabas FDT password.

Database Id Adabas Overrides the Adabas database ID in the PowerExchange data map.
Override

File Id Override Adabas Overrides the Adabas file ID in the PowerExchange data map.

148 Chapter 5: Working with Sessions

 Attribute Source Description
Name Type

Library/File DB2i5OS Overrides the library and file names in the extraction map.
Override Real Time Specify the full library name and file name in the format:
library/file
Alternatively, specify an asterisk (*) wildcard for the library name to
retrieve changes for all files of the same file name across multiple
libraries.
This attribute overrides the Library/File Override attribute on the
application connection.

Source Schema Oracle Overrides the source schema name.

Override

Filter Overrides All Filters the source data that PowerExchange reads based on specified
conditions.
PWXPC adds filter conditions specified to the WHERE clause on the
SELECT SQL statement and passes the SQL statement to
PowerExchange for processing. You can use any filter condition syntax
that PowerExchange supports for NRDB SQL. For more information, see
the PowerExchange Reference Manual.
For example, you can select records where a column called TYPE has a
value of A or D by specifying the following filter condition:
TYPE=‘A’ or TYPE=‘D’
To select change records where columns ID and ACCOUNT have
changed, you can use the DTL__CI columns by specifying the following
filter condition:
DTL__CI_ID=‘Y’ and DTL__CI_ACCOUNT=’Y’

SQL Query All Overrides the SQL query sent to PowerExchange, including any Filter
Override Overrides.

7. Click OK.

Configuring a Change or Real-time Session for a Relational Source

Configure a Change or Real-time session for a relational source from the Task Developer.

1. In the Task Developer, double-click the session.

The Edit Tasks dialog box appears.
2. Click the Properties tab and change the following fields:
• In the Commit Type field, select Source.
• Clear the Commit on End of File option.
• To enable recovery for the session, set the Recovery Strategy property to Resume from last
checkpoint. This setting is important for CDC sessions because it ensures that data and restart
tokens are properly handled.
• If you write change data to a relational target, ensure that the Session Retry on Deadlock property
is not selected in the Performance section to prevent retries of target writes when a deadlock occurs.
By default, this property is not selected. Also, in the relational connection definition for the target,
ensure that the Connection Retry Period attribute is set to the default value of 0 to prevent

Extracting CDC Data in Change and Real-time Modes 149

 database connection retries. Changing either of these settings from the default values requires
enough buffers in the DTM buffer pool to hold the largest transaction that the session processes.
If you do not use the default settings for the Session Retry on Deadlock property and Connection
Retry Period attribute, the CDC session might hang with the following message because not enough
buffers are available:
WRT_8450 : The writer is waiting for buffer blocks. To improve session
performance and to prevent the session from hanging, increase the DTM buffer pool
size in the session properties.
In this case, set the Commit Interval session property and the Maximum Rows Per commit
attribute on the source connection to the same value, for example, 1000. Although the Commit
Interval property does not control when the session commits data to targets, it does affect the
number of buffers in the DTM buffer pool that can be used to hold a transaction. Use the Maximum
Rows Per commit connection attribute to divide the rows in a large transaction into subsets for
subpacket commit processing on the target. For more information about controlling commit
processing, see “Connection Attributes for CDC Commit Processing ” on page 126.
3. Click the Sources view on the Mapping tab.
In the Readers field, select one of the following CDC readers based on the source type :
• For DB2 for z/OS, select either PowerExchange CDC Change or PowerExchange CDC Real-time for
DB2zOS.
• For DB2 for i5/OS, select either PowerExchange CDC Change or PowerExchange CDC Real-time for
DB2i5OS.
• For DB2 for Linux, UNIX, and Windows, select either PowerExchange CDC Change or
PowerExchange Real-time for DB2LUW.
• For Oracle, select PowerExchange CDC Real-time for Oracle.
• For Microsoft SQL Server, select either PowerExchange CDC Change or PowerExchange Real-time
for MSSQL.
4. In the Connection Value field, select an application connection.
Tip: For CDC Change and Real-time readers, select an application connection only for the first CDC
source. For subsequent CDC sources, select a Connection Type of None. PowerExchange group
source processing only uses the connection information from the first application connection.
Subsequent application connection specifications are not required and might cause session failures.
5. Optionally, open the application connection to override any connection values.
6. In the Properties settings, configure the attributes that apply to the source type.

150 Chapter 5: Working with Sessions

 The following table describes required and optional attributes:

Attribute Source Type Description

Name

Extraction Map All Required. The PowerExchange extraction map name for the CDC
Name source. You must specify the extraction map name for the relational
source.

Library/File DB2i5OS Real Optional. Overrides the library and file names in the extraction map.
Override Time Specify the full library name and file name in the format:
library/file
Alternatively, specify an asterisk (*) wildcard for the library name to
retrieve changes for all files of the same file name across multiple
libraries.
This attribute overrides the Library/File Override value on the
application connection.

Source Schema Oracle Change Optional. Overrides the source schema name.
Override and Real Time

7. Click OK.

Related Topics:
• “Connections” on page 61
• “Asterisk Wildcard in i5/OS Library or File Overrides” on page 151
• “Restart and Recovery” on page 163
• “Filtering Source Data with PWXPC” on page 210

Asterisk Wildcard in i5/OS Library or File Overrides

For real-time extraction, if an i5/OS file exists in many libraries with the same name and layout, you can
specify the asterisk (*) wildcard for the name in the Library/File Override attribute. PowerExchange can then
read changes for the specified file name that are recorded in the same journal, across all of the libraries that
contain the file.

With the asterisk wildcard, you configure only a single source in the mapping.

To enter the asterisk wildcard in the Library/File Override property, use the following syntax:
*/file_name

Consider the following rules and guidelines:

• You can use the asterisk wildcard in the Library/File Override property only for real-time extractions.
• DB2 for i5/OS files that are included in a mapping require PowerExchange capture registrations. By using
the asterisk as a library override, you can map and register a single file for change data capture in the
PowerExchange Navigator. CDC sessions can extract changes for all of the file instances that occur
across all libraries.
• To extract changes for multiple file instances, the files must use the same journal and be defined with
IMAGES(*BOTH).
• The specified file name must be a valid i5/OS file name. You cannot specify DB2 table names.

Extracting CDC Data in Change and Real-time Modes 151

 • If you use i5/OS-generated file names, make sure that you understand the journal contents. An i5/OS-
generated file name is composed of only the first 5 characters of the DB2 table name followed by 5 digits.
These file names might result in data extraction from unexpected files.
• Each time a real-time extraction session runs, PowerExchange checks the journal for files that match the
file name. The files for which changes are extracted can vary from session to session if you add new
libraries that contain the same file name and use the same journal.
• Restart tokens are only for sources that are included in a mapping. PWXPC indicates the presence of a
Library/File Override value for a session by prefacing the extraction map name with the specified
override.
The following example shows the restart tokens for a source that is processed by a session for which the
Library/File Override attribute specifies */file_name:
<*/file_name>d3instance.regname_tablename=restart1
<*/file_name>d3instance.regname_tablename=restart2
You do not need to include the Library/File Override information in the restart token file when overriding
a restart token. PWXPC includes the override information if it is found in the session properties.

Treating Source Rows as Updates

PowerExchange extraction maps include the DTL_CAPXACTION column, which indicates whether the
change is an insert, update, or delete. When you import an extraction map in Designer, PWXPC includes this
column in the source definition. The PowerCenter Integration Service uses the column to determine which
action PowerCenter applies to the row when the row reaches the target.

You might need to configure a session to treat each source row as an update, regardless of the value of the
DTL_CAPXACTION column. For example, suppose the change stream includes multiple row inserts that
have the same value for a non-key field in the source table. If this field is the key field of the target table, an
attempt to insert the rows into the target table results in a primary key constraint violation.

To configure a session to move each source row to the target as an update, complete the following steps:

1. In the Task Developer, double-click the session to edit it.

2. Click the Properties tab.
3. For the Treat Source Rows As attribute, select Data driven.
4. In the Mapping Designer, add an Update Strategy transformation to the mapping.
5. Open the Update Strategy transformation.
6. Click the Properties tab.
7. Click the button next to the Update Strategy Expression field.
8. Enter an update strategy expression to always mark rows as updates.
9. Validate the expression and click OK.
10. Click OK.
11. Connect the ports in the Update Strategy transformation to another transformation or a target instance.
When you run the session, PowerCenter moves each source row to the target as an update if possible. If a
row with a matching primary key value is not present in the target system, PowerCenter inserts the row.

Note: For CDC data sources, if you select Insert, Update, or Delete for the Treat Source Rows As property,
the Integration Service ignores the selection and uses the action that is indicated in the DTL_CAPXACTION
column.

152 Chapter 5: Working with Sessions

Loading Data to PowerExchange Targets
You can load data to relational targets as well as to nonrelational targets.

Related Topics:
• “PowerExchange Client for PowerCenter (PWXPC)” on page 15

Configuring Sessions to Load Data to Relational Targets

To load data to a relational target, select a relational database connection. You can then configure properties
for session as you would any other relational target.

Note: DB2 for i5/OS tables can also be accessed as flat files. If the table has been imported into
PowerExchange as a sequential data map, you can use a PWX NRDB Batch application connection to write
to it.

To configure sessions to load data to relational targets:

1. In the Task Developer, double-click the session to edit it.

2. Click the Targets view on the Mapping tab.
3. In the Writers setting, select Relational Writer to run sessions with relational targets.
4. In the Connections Value field, select a relational database connection from one of the following types:
• For DB2 for i5/OS, select PWX DB2i5OS.
• For DB2 for Linux, UNIX, and Windows, select PWX DB2LUW.
• For DB2 for z/OS, select PWX DB2zOS.
• For Microsoft SQL Server, select PWX MSSQLServer.
• For Oracle, select PWX Oracle
5. Configure any other session properties.
Note: If you select the Truncate Table Option session for a PWX DB2zOS target connection, by default
PowerExchange issues a TRUNCATE statement to truncate table contents before loading new data.
If the PWX Override connection attribute includes the DB2TRUNCASDEL=ON override,
PowerExchange issues a DELETE statement instead of a TRUNCATE statement.
The DB2TRUNCASDEL override applies only to DB2zOS target connections. For DB2i5OS target
connections, PowerExchange always issues a DELETE statement. For DB2LUW target connections,
PowerExchange always issues a TRUNCATE statement.
6. Click OK.

Configuring Sessions to Load Data to Nonrelational Targets

PWXPC can load data to nonrelational targets. Nonrelational target types include Adabas, IMS, sequential
data sets, flat files, and VSAM files.

Based on the target type, PowerCenter selects the correct PowerExchange Writer. After you select a PWX
NRDB Batch application connection, you can configure session properties for the target.

1. In the Task Developer, double-click the session to edit it.

2. On the Mapping tab, expand the Targets node in the Transformations tree. Then click a nonrelational
target.
PowerCenter sets the writer value based on the selected target type.

Loading Data to PowerExchange Targets 153

 3. Under the Connections, in Value field, select a PWX NRDB Batch application connection.
4. Under Properties, set the session attributes for the target type, as needed.
The following table describes these attributes in alphabetical order:

Attribute Name Target Type Description

ADABAS ADABAS Adabas file password.

Password If the ADABAS FDT for the target file is password protected, enter the
ADABAS FDT password.

BLKSIZE SEQ (MVS MVS data set block size.

only) Default is 0, which means use the best possible block size.

DATACLAS SEQ (MVS MVS SMS data class name.

only)

Delete SQL ADABAS, Overrides default Delete SQL sent to PowerExchange.

Override IMS, VSAM

Disp SEQ (MVS MVS data set disposition.

only) Valid values:
- OLD
- SHR
- NEW
- MOD
Default is MOD if the data set exists, and NEW if it does not.

File Name SEQ, VSAM Overrides the data set or file name in the PowerExchange data map.
Override Enter the complete data set or file name.
For i5/OS, use the following format: library_name/file_name.

IMS AM Override IMS Overrides the IMS access method in the imported data map for the
target with the other allowable access method. The session then uses
the override access method at run time.
- If you imported a target data map that specifies the DL/1 BATCH access
method, enter O to override it with the IMS ODBA access method. For
ODBA access, you must also specify the IMS PSBNAME Override and
IMS PCBNAME Override attributes.
- If you imported a target data map that specifies the IMS ODBA access
method, enter D to override it with the DL/1 BATCH access method,
which provides DL/I or BMP access. You must also specify the IMS
PCBNUMBER Override attribute.
Important: Before you run the session with an access method
override, ensure that you complete the PowerExchange configuration
tasks for the new access method. For example, if the override is DL/1
BATCH, you must configure LISTENER and NETPORT statements in
the DBMOVER member and configure the netport JCL. If the override
is IMS ODBA, you must perform other configuration tasks. For more
information, see "IMS Data Maps" in the PowerExchange Navigator
User Guide.

154 Chapter 5: Working with Sessions

Attribute Name Target Type Description

IMS PCBNAME IMS If you imported an IMS ODBA data map for the target and did not
Override override the access method, this value overrides the PCB name from
the data map. If you specified ODBA access as an override in the IMS
AM Override attribute, you must enter this value. A PCB name is
required for ODBA access.
A PCB name is not used for DL/I or BMP access.

IMS IMS If you imported a DL/1 BATCH data map for the target and did not
PCBNUMBER override the access method, this value overrides the PCB number from
Override the data map. If you specified DL/I or BMP access as an override in
the IMS AM Override attribute, you must enter this value. A PCB
number is required for DL/I or BMP access.
A PCB number is not used for IMS ODBA access.

IMS PSBNAME IMS If you imported an IMS ODBA data map for the target and did not
Override override the access method, this value overrides the PSB name from
the data map. If you specified ODBA access as an override in the IMS
AM Override attribute, you must enter this value. A PSB name is
required for ODBA access.
If you use DL/I batch or BMP access and specify this override, you
must also specify the PSB=%PSBNAME substitution variable in the
netport JCL. The override value then replaces the substitution variable
in the JCL.
If you specify the PSB=%1 substitution variable instead of
PSB=%PSBNAME in the netport JCL, the session uses the PSB name
in the NETPORT statement, if specified. In this case, you need a
separate NETPORT statement for each PSB. To avoid exceeding the
limit of ten NETPORT statements, use this override with %PSBNAME
substitution variable instead.

IMS SSID IMS If you imported an IMS ODBA data map for the target and did not
Override override the access method, use this value to override the IMS
subsystem ID (SSID). If you specified ODBA access as an override in
the IMS AM Override attribute, you must enter this value. An SSID is
required for ODBA access.
If you use the IMS DL/1 BATCH access method and a BMP netport
job, you can use this override with the %IMSID substitution variable in
the netport JCL. This override replaces the substitution variable to
specify the IMS SSID to use for the session. By using the substitution
variable and override together, you can use the same netport JCL to
access multiple IMS environments, such as development, testing, and
production environments.
Note: An IMS SSID is not required for DL/I batch access to IMS data
or for access to an IMS unload file.

Initialize Target VSAM Select this option to have PowerExchange allow both inserts and
updates into empty VSAM data sets.
If this option is not selected, PowerExchange only allows inserts into
empty VSAM data sets.

Insert Only ADABAS, Processes updates and deletes as inserts.

IMS, VSAM Note: You must select this option when the target has no keys.

Loading Data to PowerExchange Targets 155

 Attribute Name Target Type Description

Insert SQL All Overrides the default Insert SQL sent to PowerExchange.
Override

LRECL SEQ (MVS MVS data set logical record length.

only) This value is ignored if Disp is not MOD or NEW.
Default is 256.

Map Name All Overrides the target PowerExchange data map name.
Override Note: PWXPC sends the file name that is specified for the source in
the mapping unless this name is overridden in File Name Override
attribute.

MGMTCLAS SEQ (MVS SMS management class name.

only) This value is ignored if Disp is not MOD or NEW.

MODELDSCB SEQ (MVS MVS Model DSCB for non-SMS-managed GDG data sets.
only) This value is ignored if Disp is not MOD or NEW.

Post SQL All One or more SQL statements that are executed after the session runs
with the target database connection.

Pre SQL All One or more SQL statements that are executed before the session
runs with the target database connection.
Note: If you specify a SQL statement in the Pre SQL property for a
bulk data movement session that uses a PWX NRDB Batch connection
and writer partitioning, the session executes the SQL statement in
each partition. The session fails if you specify a SQL statement such
as CREATEFILE that can run only once for the session. The session
executes the statement in the first partition and then fails when trying
to execute it again in another partition. Either enter a SQL statement
that can run once in each partition, or do not specify the Pre SQL
attribute for writer partitions.

Primary Space SEQ (MVS MVS primary space allocation, in the units specified in the Space
only) attribute.
This value is ignored if Disp is not MOD or NEW.
Default is 1.

RECFM SEQ (MVS MVS record format. Valid values are F, V, FU, FB, VU, VB, FBA, and
only) VBA.
This value is ignored if DISP is not MOD or NEW.

Schema Name All Overrides the schema name in the target PowerExchange data map.
Override Note: PWXPC sends the file name for the source in the mapping
unless this name is overridden in File Name Override attribute.

Secondary SEQ (MVS MVS secondary space allocation, in the units specified in the Space
Space only) attribute.
This value is ignored if Disp is not MOD or NEW.
Default is 1.

156 Chapter 5: Working with Sessions

 Attribute Name Target Type Description

Space SEQ (MVS Type of units for expressing primary or secondary space for MVS data
only) sets. Valid values are:
- CYLINDER
- TRACK
This value is ignored if Disp is not MOD or NEW.
Default is TRACK.

STORCLAS SEQ (MVS SMS storage class name.

only) This value is ignored if Disp is not MOD or NEW.

Truncate Table IMS, VSAM Truncates, or deletes, table contents before loading new data.
Option Note: VSAM data sets must be defined with the REUSE option for this
truncate option to function correctly.

UNIT SEQ (MVS MVS unit type.

only) This value is ignored if Disp is not MOD or NEW.
Default is SYSDA.

Update SQL ADABAS, Overrides the default Update SQL that is sent to PowerExchange.
Override IMS, VSAM

Upsert ADABAS, Processes failed inserts as updates and updates as inserts.

IMS, VSAM

VOLSER SEQ (MVS MVS volume serial number.

only) This value is ignored if Disp is not MOD or NEW.

5. Click OK.

Constraint-Based Loading for Relational Targets

In the Workflow Manager, you can specify constraint-based loading for a session. When you select this
option, the PowerCenter Integration Service orders the target load on a row-by-row basis. For every row
generated by an active source, the PowerCenter Integration Service loads the corresponding transformed
row first to the primary key table and then to any foreign key tables.

Configuring Constraint-Based Loading for Relational Targets

In the Workflow Manager, you can specify constraint-based loading for a session. When you select this
option, the PowerCenter Integration Service orders the target load on a row-by-row basis. For every row
generated by an active source, the PowerCenter Integration Service loads the corresponding transformed
row first to the primary key table, then to any foreign key tables.

The Constraint based load ordering attribute applies only to insert operations. Change data normally
contains a mixture of insert, update, and delete operations. PowerCenter provides a custom property to
enable complete constraint-based loading.

When you enable complete constraint-based loading, change data is loaded to targets in the same
Transaction Control Unit (TCU) by using the row ID assigned to the data by the CDC Reader. As a result,

Constraint-Based Loading for Relational Targets 157

 data is applied to the targets in the same order in which it was applied to the sources. The following message
will be issued in the session log to indicate that this support is enabled:
WRT_8417 Complete Constraint-Based Load Ordering is enabled.

To enable complete constraint-based loading, specify FullCBLOSupport=Yes in the Custom Properties

attribute on the Config Object tab. This property can also be set in the PowerCenter Integration Service,
which makes it applicable to all workflows and sessions that use that the PowerCenter Integration Service.

If you use complete constraint-based loading, your mapping must not contain active transformations which
change the row ID generated by the CDC Reader. The following transformations change the row Id value:

• Aggregator
• Custom, configured as an active transformation
• Joiner
• Normalizer
• Rank
• Sorter

Variables for Session and Connection Properties

You can use variables to assign values to properties in the Designer and Workflow Manager and to override
some service and service process properties.

You can also use variables to override PWXPC session properties. You can also use variables to override
source or target relational or application connection attributes other than PWXPC connection attributes.

PWXPC variables must begin with $PWX.

The following table lists the PWXPC session source and target properties for which you can specify
variables:

Session Property Source or Target Type

DATACLAS MVS sequential file sources and targets

DB2 Sub System Id DB2 image copy and DB2 data map sources

DB2 Table name DB2 data map sources

Delete SQL override Nonrelational targets

Extraction Map Name Change data capture sources

File Name Override VSAM and sequential sources and targets

Filter Overrides All sources

Image Copy Dataset DB2 image copy sources

IMS Unload File Name IMS sources

Insert SQL override Nonrelational targets

158 Chapter 5: Working with Sessions

 Session Property Source or Target Type

Map Name Override All sources and nonrelational targets

MGMTCLAS MVS sequential file sources and targets

MODELDSCB MVS sequential file sources and targets

Post SQL Nonrelational targets

Pre SQL Nonrelational targets

Schema Name Override All

Source Schema Override Relational change data capture sources and DB2 image copy

Sql Query Override All sources

STORCLAS MVS sequential file sources and targets

UNIT MVS sequential file sources and targets

Unload File Name Adabas and DB2 unload file sources

Update SQL override Nonrelational targets

Workflows that Connect to a PowerExchange

Listener Service
When you run a workflow, the PowerCenter Integration Service can connect to the PowerExchange Listener
through the Listener Service. The Listener Service is an application service in the Informatica domain that
manages a PowerExchange Listener. The Listener Service can define backup nodes to run a
PowerExchange Listener process if the PowerExchange Listener on the primary node becomes unavailable.

The PowerCenter Integration Service connects to the PowerExchange Listener based on the way you
configure the NODE statement in the DBMOVER configuration file:

• If the NODE statement includes the SERVICE_NAME parameter, the PowerCenter Integration Service
connects to the Listener through the Listener Service.
• If the NODE statement does not include the SERVICE_NAME parameter, the PowerCenter Integration
Service connects directly to the Listener. It does not connect through the Listener Service.
For more information about the Listener Service, see the Informatica Administrator Guide.

Workflows that Connect to a PowerExchange Listener Service 159

Configuring PowerExchange Services for Workflows
Before you run a workflow that uses PowerExchange services, you must configure the services.

To configure PowerExchange Services for workflows:

1. Install PowerExchange on the PowerCenter Integration Service machine and on each node where you
plan to run the Listener Service and Logger Service.
2. Customize the DBMOVER file on each node where you plan to run PowerExchange Services, as follows:
• Configure the LISTENER statement to identify the PowerExchange Listener.
• For CDC sessions, configure CAPI_CONNECTION statements to control extraction processing.
Include the same statements in each DBMOVER file.
• If the PowerExchange Listener will connect to the PowerExchange Logger for Linux, UNIX, and
Windows, customize the settings required for the PowerExchange Logger.
3. Customize the DBMOVER file on each node that runs the PowerCenter Integration Service. Configure
the SERVICE_NAME parameter of the NODE statement to identify the Listener Service.
4. If the PowerExchange Listener for the data source will connect to a PowerExchange Logger, customize
the pwxccl.cfg file on each node where you plan to run the Logger Service. Configure each pwxccl.cfg
file to point to the same PowerExchange Logger files, and verify that the files are accessible to each
node.
5. Create a PowerExchange Listener Service on the primary and backup nodes for the service.
6. If the PowerExchange Listener for the data source will connect to a PowerExchange Logger, create a
PowerExchange Logger Service on the primary and backup nodes for the service.
7. Start the services in the Administrator tool.

Pipeline Partitioning
You can use pipeline partitioning at reader and writer partition points in bulk data movement sessions.
Reader partitioning supports several partition types for data sources. Writer partitioning supports pass-
through partitioning for VSAM and sequential file targets. For more information, see the PowerExchange Bulk
Data Movement Guide and the PowerCenter Advanced Workflow Guide.

You can also use pipeline partitioning for targets in CDC sessions. However, you cannot use partitioning for
sources in CDC sessions.

Commit Processing for Bulk Data Movement

Sessions
During a bulk data movement session, by default PowerExchange commits data to the target each time the
commit interval that is defined in the session properties is reached.

You might need to change how frequently PowerExchange commits data to the target. If commits are issued
too infrequently, bulk data movement sessions with a nonrelational target might fail with an out-of-memory
error.

160 Chapter 5: Working with Sessions

 To prevent or correct this problem, perform one of the following actions:

• Reduce the Commit Interval session property. The default is 10,000 rows.
• Tune the database based on the error message that you receive.
• If necessary, you can define ENABLE_AUTOCOMMIT=Y in the DBMOVER configuration file on the client
machine. Defining ENABLE_AUTOCOMMIT=Y instructs PowerExchange to commit data to the target
each time the data buffer becomes full. You set the data buffer size for a session by defining the Default
Buffer Block Size session property in the Advanced settings of the Config Object tab.

Caution: Defining ENABLE_AUTOCOMMIT=Y alters the standard commit and rollback behavior for
PowerCenter sessions. Avoid using this statement if possible.

Commit Processing for CDC Sessions

The PowerCenter Integration Service, in conjunction with PWXPC, commits data to the target based on
commit connection attributes and the commit type.

By default, the Commit Type property on the session Properties tab specifies Target. For CDC sessions, the
Integration Service always uses source-based commit processing. When you run a CDC session that
specifies target-based commit processing, the PowerCenter Integration Service automatically changes the
commit type to source-based and writes message WRT_8226 in the session log.

For CDC sessions, PWXPC ignores the Commit Interval attribute. To control commit processing, configure
attributes on the PWX CDC Change and Real Time application connections.

Related Topics:
• “Connection Attributes for CDC Commit Processing ” on page 126

Considerations for Pre- and Post-Session SQL

Commands
You can specify pre- and post-session SQL in the Source Qualifier transformation and the target instance
when you create a mapping. When you create a Session task in the Workflow Manager you can override the
SQL commands on the Mapping tab.

For the following databases, if you include a call to a stored procedure in the pre- or post-session SQL,
PowerExchange issues error message PWX-00183:

• DB2 for Linux, UNIX, and Windows

• Microsoft SQL Server
• Oracle

Commit Processing for CDC Sessions 161

Preserving Low Values in Source Character Fields
If a data source includes character fields with hexadecimal '0' values, called low values, you can configure
PowerExchange and PowerCenter to preserve these values so that a PowerCenter session can write the low
values to a VSAM target on z/OS or to a sequential file target on Linux, UNIX, Windows, or z/OS.

To preserve low values, perform the following configuration tasks:

• Configure a PowerCenter session that uses the PowerExchange Client for PowerCenter (PWXPC) instead
of the PowerExchange ODBC interface.
• If you want to write low values from a nonrelational data source on z/OS to a nonrelational target on z/OS
without any translation of the values, complete the following tasks in PowerCenter:
- On the Config Object tab for the session, enter PreserveLowValues=Yes in the Custom Properties
field.
- In the source and target PWX NRDB Batch application connections that are included in the PowerCenter
workflow, ensure that the Convert character data to string option is not selected.
- In the DBMOVER configuration files on the source system and the Integration Service system, set the
LOWVALUES statement to Y. For more information, see the PowerExchange Reference Manual.
If you do not perform these configuration tasks, PowerExchange does not preserve low values. It interprets a
hexadecimal '0' value as the end of a column string and pads the remaining length of the string with spaces.
The portion of the column string after the hexadecimal '0' value is not written to the target.

162 Chapter 5: Working with Sessions

CHAPTER 6

Restart and Recovery

This chapter includes the following topics:

• Restart and Recovery Overview, 163

• Understanding PWXPC Restart and Recovery, 164
• Creating Recovery Tables, 171
• Configuring the Restart Token File, 173
• PWXPC Restart and Recovery Operation, 177

Restart and Recovery Overview

PWXPC supports restart and recovery processing. Learn how to configure your CDC sessions to use this
processing.

Each source in a CDC session has unique restart information, referred to as restart tokens. PWXPC
manages the CDC restart information. The PowerCenter Integration Service provides recovery for the target
files and tables in CDC sessions.

To extract change data from the change stream, PWXPC provides restart information for CDC sources to
PowerExchange. PowerExchange reads the change stream on the source platform and provides complete
units of work (UOWs) to PWXPC. A UOW is a collection of changes within a single commit scope, which are
made by a transaction on the source system. Based on the commit interval information that is specified on
the CDC session connection, PWXPC periodically flushes complete UOWs to the PowerCenter Integration
Service.

Target recovery and restart information is stored as the target tables and files are updated by the
PowerCenter Integration Service. The PowerCenter Integration Service and PWXPC use this information to
recover and restart stopped or failed sessions from the point of interruption.

Related Topics:
• “CDC Application Connection Attributes” on page 116
• “Commit Processing for CDC Sessions” on page 161
• “Understanding PWXPC Restart and Recovery” on page 164
• “Configuring the Restart Token File” on page 173
• “Configuring CDC Sessions” on page 178
• “PWXPC Restart and Recovery Operation” on page 177
• “Enabling Session Recovery for a CDC Session” on page 177

163
 • “Extracting CDC Data in Change and Real-time Modes” on page 146

Understanding PWXPC Restart and Recovery

PowerCenter and PWXPC provide restart and recovery functionality.

Session Recovery
PowerCenter and PWXPC restart and recovery functionality are used when you configure the Recovery
Strategy Resume from the last checkpoint. Enabling a resume recovery strategy ensures that both the target
data and CDC restart information can be recovered in the event of session failure.

Warning: Do not enable recovery processing if any of the targets in the CDC session use the File Writer to
write CDC data to flat files. The restart tokens for all targets in the session, including relational targets, will be
compromised if there is a flat file target in the same session. Data loss or duplication may occur.

When you enable a resume recovery strategy, the PowerCenter Integration Service provides recovery for the
target tables and files and PWXPC provides recovery for the CDC restart information. PWXPC issues the
following message indicating that recovery is in effect:
PWXPC_12094 [INFO] [CDCRestart] Advanced GMD recovery in affect. Recovery is automatic

The PowerCenter Integration Service stores the session state of operation in the shared location,
$PMStorageDir. The PowerCenter Integration Service saves relational target recovery in the target database.

CDC restart information, also called restart tokens, originates from PowerExchange on the CDC source
platform. PWXPC stores CDC restart information in different locations based upon the target type:

• For nonrelational targets, PWXPC stores the CDC restart information in the shared location,
$PMStorageDir, in state files on the Integration Service platform.
• For relational targets, PWXPC stores the CDC restart information in state tables in the target database.
When the PowerCenter Integration Service performs recovery, it restores the state of operation to recover the
session from the point of interruption. It uses the target recovery data to determine how to recover the target
tables. PWXPC and PowerExchange use the CDC restart information to determine the correct point in the
change stream from which to restart the extraction.

Recovery Tables
For relational targets, the PowerCenter Integration Service creates the following recovery tables in the target
database:

• PM_RECOVERY. This table contains target load information for the session run. The PowerCenter
Integration Service removes the information from this table after each successful session and initializes
the information at the beginning of subsequent sessions.
• PM_TGT_RUN_ID. This table contains information the Integration Service uses to identify each target on
the database. The information remains in the table between session runs. If you manually create this
table, you must create a row and enter a value other than zero for LAST_TGT_RUN_ID to ensure that the
session recovers successfully.
• PM_REC_STATE. This table contains restart information for CDC sessions. The restart information
recorded in the table contains the application name and restart tokens for the session. The restart
information remains in the table permanently. The PowerCenter Integration Service updates it with each
commit to the target tables.

164 Chapter 6: Restart and Recovery

 If you edit or drop the recovery tables before you recover a session, the PowerCenter Integration Service
cannot recover the session. If you disable recovery, the PowerCenter Integration Service does not remove
the recovery tables from the target database. You must manually remove the recovery tables.

If you want the PowerCenter Integration Service to create the recovery tables, grant table creation privilege
to the database user name for the target database connection.

If you do not want the PowerCenter Integration Service to create the recovery tables, create the recovery
tables manually.

Tip: If you use PowerExchange relational target connections, you must manually create these tables so you
can assign the desired database attributes.

Related Topics:
• “Recovery Table Creation with PowerExchange Targets” on page 172
• “Creating the Recovery Tables Manually” on page 173

Recovery State Table

The recovery state table contains the state and CDC restart information for a CDC session. The table resides
in the same target database as target tables. The PowerCenter Integration Service creates the state table
automatically if it does not exist.

The PowerCenter Integration Service creates an entry in the state table for each CDC session. These entries
may comprise more than one row. CDC sessions with heterogeneous relational target tables have entries in
the state table in each unique target database. For example, a CDC session which targets Oracle and SQL
Server tables has an entry in the state table in the target Oracle database instance and in the target SQL
Server database.

Each session entry in the state table contains a number of repository identifiers as well as execution state
data like the checkpoint number and the CDC restart information. The following columns contain PWXPC-
specific restart information:

• The APPL_ID column contains the application name specified in the source application connection, with
the task instance ID appended.
• The STATE_DATA column, which contains the restart tokens for the session, is a variable 1,024-byte
binary column. If the number of restart tokens for a session causes the data to exceed 1,024 in length,
additional rows are added to accommodate the remainder of the restart information. The SEQ_NUM field
is increased by one, starting from zero, for each additional row added for a session entry.
The majority of the columns in the table are task and workflow repository attributes. These repository
attributes remain static unless the task or workflow is altered. The following examples are actions that alter
these repository attributes:

• Adding or removing sources or targets from the mapping used by the session
• Moving the workflow or session to a different folder
• Moving the session to a different workflow
The PowerCenter Integration Service uses the APPL_ID column to retrieve CDC restart information. Since
the APPL_ID column contains the application name and the task instance ID for the session, adding and
removing sources or targets from the session affects restart processing.

During session initialization, the PowerCenter Integration Service reads the state table looking for an entry
that matches APPL_ID information for the CDC session. If a match is found, the PowerCenter Integration
Service uses that entry for target recovery processing. PWXPC uses the CDC restart information stored in
the STATE_DATA column to perform restart and recovery processing.

Understanding PWXPC Restart and Recovery 165

 Related Topics:
• “Creating Recovery Tables” on page 171
• “Changing CDC Sessions and Restarting from the Point of Interruption” on page 183

Recovery State File

The recovery state file is similar to the state table. PWXPC uses the state file to store the CDC restart
information for nonrelational target files. Nonrelational target files include MQ Series message queues,
PowerExchange nonrelational targets, and other PowerCenter nonrelational targets. The PowerCenter
Integration Service automatically creates the recovery state file in the shared location, $PMStorageDir, if it
does not exist. The file name for the recovery state file is prefixed with:
pm_rec_state_application_name

The PowerCenter Integration Service uses the application name from the source CDC connection for the
application name value in the state file name prefix. The PowerCenter Integration Service includes the
complete file name in message CMN_65003. The remainder of the fields in the file name are task and
workflow repository attributes. These repository attributes remain static unless the task or workflow is altered.
The following examples are actions that alter these repository attributes:

• Adding or removing sources or targets from the mapping used by the session
• Moving the workflow or session to a different folder
• Moving the session to a different workflow

Related Topics:
• “Changing CDC Sessions and Restarting from the Point of Interruption” on page 183

Restart Token File

PWXPC stores the restart tokens in the state table in the target database for relational targets and in the
state file on the PowerCenter Integration Service platform for nonrelational targets. The restart token file
provides:

• Initial restart tokens for new CDC sessions

• Overrides for the restart tokens in the state table or file for existing CDC sessions
PWXPC uses the restart token file in the folder specified in the RestartToken File Folder attribute of the
source CDC connection. PWXPC automatically creates this folder, if it does not exist, when the attribute
contains the default value of $PMRootDir/Restart. PWXPC does not automatically create any other restart
token folder name.

During session initialization, PWXPC:

• Uses the name specified in the RestartToken File Name attribute to create an empty restart token file, if
one does not already exist.
Creates a merged view of the restart tokens by reconciling the restart tokens specified in the restart token
file with those in state tables and the state file for all relational and non-relations targets, respectively.
• Places the results of the restart token reconciliation process into an initialization file in the restart token
file directory and empties out the restart token file.
Emptying the restart token file ensures that it does not override the state table or state file restart tokens
with the same restart information the next time the session is run.

166 Chapter 6: Restart and Recovery

 During normal termination, PWXPC writes the ending restart tokens into a termination file in the restart token
file directory. The restart token files containing the initialization and termination restart tokens have the
following names:
<restart_token_file_name>yyyymmddhhmmss_init
<restart_token_file_name>yyyymmddhhmmss_term

Where:

• restart_token_file_name is the restart token file name from the CDC connection
• yyyymmddhhmmss is the initialization file creation time stamp
• init or term is for initialization and termination files, respectively
For example, a CDC source application connection specifies a restart token file name of my.app.txt, which
does not exist. PWXPC creates the following files on the PowerCenter Integration Service platform in the
restart token file folder specified in the connection:

• my.app.txt
• my.app.txtyyyymmddhhmmss_init
• my.app.txtyyyymmddhhmmss_term
The restart token file, my.app.txt, is empty. The timestamps on both the initialization and termination files are
the same to indicate that they are related to the same run. The termination file may not exist or may be empty
if the session fails.

If you use the default value of zero for the connection attribute Number of Runs to Keep RestartToken File,
PWXPC keeps only one copy of the paired initialization and termination files. Otherwise, PWXPC uses the
value specified in that attribute to determine the number of backup copies of these paired files to keep.
During termination, PWXPC removes any additional pairs of the backup files beyond the Number of Runs to
Keep RestartToken File value.

Related Topics:
• “Determining the Restart Point” on page 167

Determining the Restart Point

Start CDC sessions in one of the following ways:

• Cold start. When you cold start a CDC session, PWXPC reads only the restart token file to acquire restart
tokens for all sources and makes no attempt to recover the session. The session continues to run until
stopped or interrupted.
• Warm start. When you warm start a CDC session, PWXPC reconciles the restart tokens provided in the
restart token file, if any, with any restart tokens that exist in the state file or state tables. If necessary,
PWXPC performs recovery processing. The session continues to run until stopped or interrupted.
• Recover operation. When you recover a CDC session, PWXPC reads the restart tokens from the state
file and state tables and writes them into the restart token file. If necessary, PWXPC performs recovery
processing. After PWXPC finishes updating the restart token file and doing any necessary recovery, the
session ends.
Each CDC source in the CDC session has its own unique restart point. You should create and populate the
restart token file with restart points for each source prior to running a CDC session for the first time. If you do
not provide restart tokens in the restart token file and no existing entry exists for the session in the state
tables or the state file, then PWXPC passes null restart tokens to PowerExchange for all sources in the
session.

The restart tokens PWXPC uses varies based on whether you warm or cold start the CDC session and
whether you provide any overriding restart tokens in the restart token file.

Understanding PWXPC Restart and Recovery 167

 Related Topics:
• “Starting CDC Sessions” on page 180

Initiating Cold Start Processing

You can cold start workflows and sessions with the cold start command from Workflow Manager, Workflow
Monitor, or pmcmd. PWXPC uses the restart tokens provided in the restart token file, disregarding the state
file and state tables, as follows:

Restart Token file is empty or does not exist.

PWXPC assigns null restart tokens to all sources in the session.

Restart token file contains explicit override statements.

PWXPC assigns the restart tokens supplied in the restart token file to the specified sources. PWXPC
assigns the oldest restart point of the restart tokens specified to all remaining sources.

Restart token file contains special override statement.

PWXPC assigns the restart tokens supplied in the restart token file to all sources.

Restart token file contains special override statement and explicit override statements.

PWXPC assigns the restart tokens supplied in the restart token file in the explicit override statements to
the specified sources. PWXPC assigns the restart tokens supplied in the special override statement to all
remaining sources.

Initiating Warm Start Processing

You can warm start workflows and sessions with the start or restart commands from Workflow Manager,
Workflow Monitor, or pmcmd. PWXPC merges the restart tokens provided in the restart token file with any
restart information that exists in the state file and the state tables as follows:

Restart token file is empty or does not exist.

• No state file (nonrelational target) or no entry in a state table (relational target) for the session.

PWXPC assigns null restart tokens to all sources in the session.

• If state file (nonrelational target) or entry in a state table (relational target) exists for some but not all
sources in the session:

PWXPC assigns the restart tokens found in the state file or state tables to the appropriate sources.
PWXPC assigns the oldest restart point of the restart tokens available to all remaining source without
restart tokens.

• If state file (nonrelational target) or entry in a state table (relational target) exists for all sources in the
session:

PWXPC uses the restart tokens from the state file or state tables.

Restart token file contains explicit override statements.

• If no state file (nonrelational target) or no entry in a state table (relational target) for the session:

PWXPC assigns the restart tokens supplied in the restart token file to the specified sources. PWXPC
assigns the oldest restart point of the restart tokens specified in the restart token file to all remaining
sources without restart tokens.

• If state file (nonrelational target) or entry in a state table (relational target) exists for some but not all
sources in the session:

168 Chapter 6: Restart and Recovery

 PWXPC assigns the restart tokens supplied in the restart token file to the specified sources. PWXPC
assigns the restart tokens found in the state file or state tables to the appropriate sources provided they
have not been supplied in the restart token file. PWXPC assigns the oldest restart point of the restart
tokens available to all remaining sources without restart tokens.

• If state file (nonrelational target) or entry in a state table (relational target) exists for all sources in the
session:

PWXPC assigns the restart tokens supplied in the restart token file to the specified sources in the
session. PWXPC assigns the restart tokens from the state file or state tables to all remaining sources
without restart tokens.

Restart token file contains special override statement.

PWXPC assigns the restart tokens supplied in the special override statement in the restart token file to
all sources.

Restart token file contains special override statement and explicit override statements.

PWXPC assigns the restart tokens supplied in the restart token file in the explicit override statements to
the specified sources. PWXPC assigns the restart tokens supplied in the special override statement to all
remaining sources without restart tokens.

Default Restart Points

If you start a new CDC session and no restart token file exists, the PowerCenter Integration Service still runs
the session. PWXPC passes null restart tokens for all of the sources to PowerExchange.

To indicate that the restart tokens are null, PWXPC issues message PWXPC_12060. PowerExchange then
assigns the earliest possible restart point to each source.

Warning: Null restart tokens might produce unexpected results. Informatica recommends that you always
specify valid restart tokens that indicate a point in the change stream where the source and target are in a
consistent state.

The following table describes the default restart points that PowerExchange uses when it receives null restart
tokens, by source and connection type:

Source Type PWX CDC Change Connection 1 PWX Real Time Connection 2

All z/OS data sources The oldest PowerExchange The PowerExchange Logger for MVS
Condense condense file that is determines the best available restart
recorded in the CDCT file. point.
Usually, this point is the oldest restart
point for which an archive log is
available, or if no archive log is
available, the oldest restart point for an
active log.

DB2 for i5/OS The oldest PowerExchange The oldest journal receiver that is still on
Condense condense file that is the journal receiver chain.
recorded in the CDCT file.

DB2 for Linux, UNIX, and The oldest PowerExchange Logger The current log position at the time the
Windows for Linux, UNIX, and Windows log PowerExchange capture catalog was
file that is recorded in the CDCT file. created.

Understanding PWXPC Restart and Recovery 169

 Source Type PWX CDC Change Connection 1 PWX Real Time Connection 2

Microsoft SQL Server The oldest PowerExchange Logger The oldest data that is available in the
for Linux, UNIX, and Windows log Publication database.
file that is recorded in the CDCT file.

Oracle The oldest PowerExchange Logger The earliest available point in the
for Linux, UNIX, and Windows log change stream:
file that is recorded in the CDCT file - For PowerExchange Oracle CDC with
- for both PowerExchange Express LogMiner, the most recent Oracle
CDC for Oracle and PowerExchange catalog dump in the archive logs.
Oracle CDC with LogMiner. - For PowerExchange Express CDC for
Oracle, the beginning of the most recent
archive log.

Any data source type that is The oldest PowerExchange Logger The oldest PowerExchange Logger for
logged to a PowerExchange for Linux, UNIX, and Windows log Linux, UNIX, and Windows log file that
Logger for Linux, UNIX, and file that is recorded in the CDCT file. is recorded in the CDCT file.
Windows3

1. With a PWX CDC Change connection, you must use batch extraction mode.
2. With a PWX CDC Real Time connection, use real time extraction mode to extract data from the source change
stream or use continuous extraction mode to extract data that is logged to a PowerExchange Logger for Linux,
UNIX, and Windows instance.
3. Includes Linux, UNIX, and Windows data sources that are logged to a local or remote PowerExchange Logger
for Linux, UNIX, and Windows instance. Also includes z/OS and i5/OS data sources that are logged to a remote
PowerExchange Logger for Linux, UNIX, and Windows instance.

PowerExchange uses the default restart point only if all sources in a CDC session have null restart tokens. If
some sources have non-null restart tokens, PWXPC assigns the oldest restart point of those restart tokens to
sources for which no restart tokens are specified.

For example, a new CDC session contains the sources A, B, and C. The restart token file contains restart
tokens for sources A and B. The restart point for source A is older than that for source B. Source C does not
have existing or supplied restart tokens. Because some sources in the CDC session have explicit restart
points, PWXPC does not assign null restart tokens to source C. Instead, PWXPC assigns the restart point for
source A to source C because this restart point is the oldest one supplied.

Related Topics:
• “Previewing PowerExchange Change Data” on page 50

Initializing and Running CDC Sessions

After the restart information for every source is determined, PWXPC:

• Flushes the restart tokens to the state tables for relational targets and to the state file for nonrelational
targets
• Writes an empty restart token file
• Creates the initialization restart token file containing the reconciled restart information
PWXPC passes the restart tokens for all sources to PowerExchange. PowerExchange uses the oldest restart
token passed by PWXPC to start extracting data from the change stream. PowerExchange does not pass
data for a source until its restart point is reached. This prevents targets from being updated with records
processed in previous extraction runs.

170 Chapter 6: Restart and Recovery

 PWXPC continually updates the restart tokens for each source in the state table or the state file as it issues
flushes target data. With relational target tables in the same database, the PowerCenter Integration Service
updates both the target tables and the restart tokens within a single commit. The PowerCenter Integration
Service does separate commits for each unique relational database. With heterogeneous targets, the restart
tokens in one relational database may differ from those in another relational database at specific points in
time.

When you use nonrelational targets, the state file and the targets likely exist on completely different
machines. With nonrelational targets, the PowerCenter Integration Service updates the targets and the state
file in separate operations. If the session fails after the PowerCenter Integration Service commits data to the
target but before it updates the restart tokens in the state file, targets may receive duplicate data when
restarted. On warm start, PWXPC uses the last restart tokens written prior to the failure. As a result, PWXPC
re-sends data which has already been applied to the nonrelational targets.

Ending CDC Sessions

Stop CDC sessions with PowerCenter and PowerExchange commands. Sessions can also fail or terminate
unexpectedly.

When you stop a CDC session, PWXPC flushes any complete units of work that remain in its buffer. PWXPC
then waits for confirmation from the PowerCenter Integration Service that the flushed data has been written
to the targets. If it has, PWXPC writes the termination restart token file with the final restart tokens and issues
the following message:
PWXPC_12075 [INFO] [CDCRestart] Session complete. Next session will restart at: :
Restart 1 [restart1] : Restart 2 [restart2]

The PowerCenter Integration Service commits the flushed data to the targets, including the restart tokens for
relational targets. After the PowerCenter Integration Service writes the flushed data to any nonrelational
targets, it updates the state file with the restart tokens.

If the session fails, the PowerCenter Integration Service rolls back any uncommitted data and the related
restart tokens for relational targets. This leaves only the last successfully committed UOW data and restart
tokens in the relational target tables. The PowerCenter Integration Service uses relational database rollback
capabilities to ensure that uncommitted data is removed during session termination. Consistency between the
restart tokens and the relational target data is guaranteed because they are both committed within the same
commit scope.

The PowerCenter Integration Service does not do rollback processing for nonrelational targets. As a result,
duplicate data can occur on restart. You should account for this in your CDC session design.

Tip: To avoid the possibility of duplicate data, design the CDC sessions in your application to use relational
targets only.

Related Topics:
• “Stopping CDC Sessions” on page 182

Creating Recovery Tables

The PowerCenter Integration Service creates the recovery tables if they do not exist. The PowerCenter
Integration Service uses the database user name specified in the connection to create the tables. The
database user name PowerExchange uses to create the recovery tables varies based on the PWXPC
connection.

Creating Recovery Tables 171

 Note: The PowerCenter Integration Service does not create recovery queues. You must create the recovery
queue manually.

Recovery Table Creation with PowerExchange Targets

PowerExchange create recovery tables for PWX DB2zOS connections.

PWX DB2zOS Relational Connections for DB2 for z/OS Targets

When PowerExchange creates the recovery tables for PWX DB2zOS connections, it uses:

• The default DB2 database (DSNDB04).

• The PowerExchange Listener user ID if the PowerExchange Listener is running with SECURITY=0 or
SECURITY=1. This user must be granted the appropriate table creation privilege.
• The PowerExchange Listener user ID if the PowerExchange Listener is running with SECURITY=2 and
MVSDB2AF=CAF. This user must be granted the appropriate table creation privilege.
• The database user name in the target connection if the PowerExchange Listener is running with
SECURITY=2 and MVSDB2AF=RRSAF. This user must be granted the appropriate table creation
privilege.

PWX DB2i5OS Relational Connections for DB2 for i5/OS Targets

When PowerExchange creates the recovery tables for PWX DB2i5OS connections, it uses:

• The PowerExchange Listener user ID if it is running with SECURITY=0 or SECURITY=1. This user must
be granted the appropriate table creation privilege.
• The database user name in the target connection is if the PowerExchange Listener is running with
SECURITY=2. This user must be granted the appropriate table creation privilege.
• The default journal so it must be enabled for the user name.

PWX DB2LUW Relational Connections for DB2 for Linux, UNIX, and Windows
Targets
When PowerExchange creates the recovery tables for PWX DB2LUW connections, it uses:

• The default table space for user-defined tables.

• The database user name in the target connection. This user must be granted the appropriate table
creation privilege.

PWX MSSQLServer Relational Connections for Microsoft SQL Server Targets

When PowerExchange creates the recovery tables for PWX MSSQLServer connections, it uses the database
user name in the target connection. This user must be granted the appropriate table creation privilege.

PWX Oracle Relational Connections for Oracle Targets

When PowerExchange creates the recovery tables for PWX Oracle connections, it uses the database user
name in the target connection. This user must be granted the appropriate table creation privilege.

172 Chapter 6: Restart and Recovery

 Creating the Recovery Tables Manually
Consider manually creating the recovery tables so you can control attributes such as the database, table
space, and buffer pools. Informatica provides SQL scripts in the following directory:
<PowerCenter installation_directory>\server\bin\RecoverySQL

The following table lists the scripts that you can run to create the recovery tables in the target database:

Script Database

create_schema_db2.sql DB2

create_schema_ora.sql Oracle

create_schema_sql.sql SQL Server

The scripts are generic DDL. Make the appropriate changes for your environment.

Configuring the Restart Token File

You can configure the restart token file to specify the point from which you want to extract source data. If you
do not know the restart token file name and location, use one of the following options:

• Look at the PWXPC_12057 message in the session log. PWXPC includes the restart token file folder and
the restart token file name in this message.
• Open the application connection associated with the source. The application connection contains the
restart token file name and folder location. This file name overrides the file name you specified in the
application connection.
• If the restart token file name is not specified in the application connection, PWXPC uses the application
name, if specified. Otherwise, PWXPC uses the workflow name.

Warning: The Restart Token File Name must be unique for every session. Non-unique names cause
unpredictable results including session failures and potential data loss.

Restart Token File Control Statement Syntax

In the restart token file, you can specify explicit override statements, special override statements, and
comments.

For explicit override control statements, use the following syntax:

extraction_map_name={sequence_token|CURRENT_RESTART}
extraction_map_name={restart_token|CURRENT_RESTART}

For special override control statements, use the following syntax:

RESTART1={sequence_token|CURRENT_RESTART}
RESTART2={restart_token|CURRENT_RESTART}

For comments, use the following syntax:

<!-- comments

Configuring the Restart Token File 173

 The following rules and guidelines apply:

• Control statements can begin in any column.

• All control statements are optional.
• Do not include blank lines between control statements.
• Comment lines must begin with:
<!--
• You can specify both explicit override statements and one special override statement.
• An explicit override statement for a source takes precedence over any special override statement.
• On warm start, explicit override statements and the special override statement take precedence over any
restart token values stored for sources in the start table or file.

Related Topics:
• “Determining the Restart Point” on page 167

Restart Token File Control Statement Parameters

You can specify control statements in the restart token file.

Comment Statement
You can use the comment statement anywhere in the restart token file. Comment statements must begin with
“<!--”.

Explicit Override Statement

The explicit override control statement defines the restart token pair for a specific source.

A source can have multiple extraction maps, each with a different name.

You can use explicit override statements with a special override statement. The explicit override statements
provide restart tokens for specific sources, and the special override statement provides restart tokens for any
other sources.

The explicit override statement has the following syntax:

extraction_map_name={restart1_token|CURRENT_RESTART}
extraction_map_name={restart2_token|CURRENT_RESTART}

The parameters are:

extraction_ map_name

An extraction map name that matches a source in the CDC session. To determine the extraction map
name, check one of the following:

• The Extraction Map Name attribute in the session properties for relational sources.
• For CDC data map sources, the Schema Name Override and Map Name Override attributes in the
session properties. These attributes override the schema and map names in the extraction map for
the source.
• For CDC data map sources, the Schema Name and Map Name values in the source Metadata
Extensions in Designer.

174 Chapter 6: Restart and Recovery

restart1_token

The sequence token. The token format varies based on the source type. You can get the token value in
the following ways:

• Use the PRINT statement of the DTLUAPPL utility to get a hexadecimal sequence value and then
add eight zeroes to the end of that value.
• View the PowerExchange-generated DTL__CAPXRESTART1 field in extracted change records.
• Look for the sequence token in PowerExchange messages such as PWX-04565 and PWX-09959.
• Look for Restart Token 1 in PWXPC messages such as PWXPC_12060 and PWXPC_12068.

restart2_token

The restart token. The token format varies based on the source type. You can get the token value in the
following ways:

• Use the PRINT statement of the DTLUAPPL utility to get the Restart value.
• View the PowerExchange-generated DTL__CAPXRESTART2 field in extracted change records.
• Look for the Logger token in PowerExchange messages such as PWX-04565 and PWX-09959.
• Look for Restart Token 2 in PWXPC messages such as PWXPC_12060 and PWXPC_12068.

CURRENT_RESTART

Generates a current restart token. The CDC reader opens a separate connection to PowerExchange to
request generation of current restart tokens and then provides the generated restart tokens to all
applicable sources. Do not specify both CURRENT_RESTART and specific sequence and restart token
values with an extraction name.

Note: You can use CURRENT_RESTART only for real-time connections.

You can also generate current restart tokens in the Database Row Test dialog box in the
PowerExchange Navigator.

Special Override Statement

Use a special override statement to define restart tokens for one or more sources.

You can use a special override statement with explicit override statements. The special override statement
then provides restart tokens for the sources without explicit override statements or an entry in the state table
or file.

A special override statement must include both the RESTART1 and RESTART2 parameters:

RESTART1={restart1_token | CURRENT_RESTART}

Specifies the sequence token. The token format varies based on the source type. Enter one of the
following keywords:

restart1_token

A specific sequence token. You can get the token value in the following ways:

• Use the PRINT statement of the DTLUAPPL utility to a hexadecimal sequence value and then
add eight zeroes at the end.
• View the PowerExchange-generated DTL__CAPXRESTART1 field in extracted change records.
• Look for the sequence token in PowerExchange messages such as PWX-04565 and
PWX-09959.
• Look for Restart Token 1 in PWXPC messages such as PWXPC_12060 and PWXPC_12068.

Configuring the Restart Token File 175

 CURRENT_RESTART

PowerExchange generates a current restart token. The CDC reader opens a separate connection to
PowerExchange to request generation of current restart tokens and then provides the generated
restart tokens to all applicable sources.

Note: You can use CURRENT_RESTART only for real-time connections. You can also generate
current restart tokens in the Database Row Test dialog box in the PowerExchange Navigator.

RESTART2={restart2_token | CURRENT_RESTART}

Specifies the restart token. The token format varies based on the source type. Enter one of the following
keywords:

restart2_token

A specific restart token. You can get the token value in the following ways:

• Use the PRINT statement of the DTLUAPPL utility to get the Restart value.
• View the PowerExchange-generated DTL__CAPXRESTART2 field in extracted change records.
• Look for the Logger token in PowerExchange messages such as PWX-04565 and PWX-09959.
• Look for the Restart Token 2 in PWXPC messages such as PWXPC_12060 and PWXPC_12068.

CURRENT_RESTART

Generates a current restart token. The CDC reader opens a separate connection to PowerExchange
to request generation of current restart tokens and then provides the generated restart tokens to all
applicable sources.

Note: You can use CURRENT_RESTART only for real-time connections.

You can also generate current restart tokens in the Database Row Test dialog box in the
PowerExchange Navigator.

Restart Token File Example

An example restart token file demonstrates the use of explicit and special override statements to provide
restart tokens for a CDC session with seven source tables.

The example restart token file contains the following statements:

<!-- Restart Tokens for existing tables -->
restart1=000000AD775600000000000000AD77560000000000000000
Restart2=C1E4E2D34040000000AD5F2C00000000
<!-- Restart Tokens for the Table: rrtb0001_RRTB_SRC_001 -->
d1dsn7.rrtb0001_RRTB_SRC_001=0000060D1DB2000000000000060D1DB20000000000000000
d1dsn7.rrtb0001_RRTB_SRC_001=C1E4E2D340400000013FF36200000000
<!-- Restart Tokens for the Table: rrtb0001_RRTB_SRC_002 -->
d1dsn7.rrtb0002_RRTB_SRC_002=000000A3719500000000000000A371950000000000000000
d1dsn7.rrtb0002_RRTB_SRC_002=C1E4E2D34040000000968FC600000000
<!-- Restart Tokens for the Table: rrtb0001_RRTB_SRC_004 -->
d1dsn7.rrtb0004_RRTB_SRC_004=000006D84E7800000000000006D84E780000000000000000
d1dsn7.rrtb0004_RRTB_SRC_004=C1E4E2D340400000060D1E6100000000

This file contains explicit override statements for three sources: RRTB_SRC_001, RRTB_SRC_002, and
RRTB_SRC_004. The file also contains a special override statement for the remainder of the sources in the
session. When the session runs, PWXPC writes message PWXPC_12060 to the session log. The message
contains the following information:
===============================
Session restart information:
===============================
Extraction Map Name Restart Token 1 Restart Token 2 Source
d1dsn7.rrtb0001_RRTB_SRC_001 0000060D1DB2000000000000060D1DB20000000000000000 C1E4E2D340400000013FF36200000000 Restart file
d1dsn7.rrtb0002_RRTB_SRC_002 000000A3719500000000000000A371950000000000000000 C1E4E2D34040000000968FC600000000 Restart file
d1dsn7.rrtb0003_RRTB_SRC_003 000000AD775600000000000000AD77560000000000000000 C1E4E2D34040000000AD5F2C00000000 Restart file (special override)
d1dsn7.rrtb0004_RRTB_SRC_004 000006D84E7800000000000006D84E780000000000000000 C1E4E2D340400000060D1E6100000000 Restart file

176 Chapter 6: Restart and Recovery

d1dsn7.rrtb0005_RRTB_SRC_005 000000AD775600000000000000AD77560000000000000000 C1E4E2D34040000000AD5F2C00000000 Restart file (special override)
d1dsn7.rrtb0006_RRTB_SRC_006 000000AD775600000000000000AD77560000000000000000 C1E4E2D34040000000AD5F2C00000000 Restart file (special override)
d1dsn7.rrtb0007_RRTB_SRC_007 000000AD775600000000000000AD77560000000000000000 C1E4E2D34040000000AD5F2C00000000 Restart file (special override)

For sources that have explicit overrides, the message displays “Restart file” in the Source column. For
sources that have the special override restart tokens, the message displays "(special override)."

PWXPC Restart and Recovery Operation

Enable a resume recovery strategy for CDC sessions. When you enable a resume recovery strategy, the
PowerCenter Integration Service provides recovery for the target table data and PWXPC provides recovery
for the CDC restart information.

PowerExchange on the source platform provides the CDC restart information, which consists of restart
tokens describing a specific start point in the change stream. PWXPC stores the restart tokens for relational
targets in recovery state tables in the target databases. For nonrelational targets, PWXPC stores the restart
tokens in the recovery state file on the PowerCenter Integration Service machine.

PWXPC issues the following message when you specify a resume recovery strategy:
PWXPC_12094 [INFO] [CDCRestart] Advanced GMD recovery in affect. Recovery is automatic

PWXPC automatically recovers warm started sessions when a resume recovery strategy is specified.

Related Topics:
• “Understanding PWXPC Restart and Recovery” on page 164

Enabling Session Recovery for a CDC Session

Use the following procedure to enable recovery for a CDC session:

1. Select Resume from last checkpoint for the Recovery Strategy in the Properties tab.
This is the only recovery strategy that enables PWXPC and the PowerCenter Integration Service to
recover CDC sessions.
2. Change the Commit Type attribute from Target to Source.
CDC sessions always use source-based commit processing regardless of the Commit Type attribute
setting.
3. Disable the Commit at End of File attribute in the Properties tab.
The PowerCenter Integration Service, after PWXPC has shutdown, issues a commit when the session
ends. Data written to the targets after PWXPC shuts down is not reflected in the restart tokens. Disabling
this flag ensures that PWXPC issues all commits.
Warning: If you use the File Writer to write CDC data to flat files, do not enable recovery processing.
Data loss or duplication may occur since the restart tokens for all targets, including relational targets, are
compromised if there is a flat file target in the same session.
If you run a session with resume recovery strategy and the session fails, do not alter the mapping, the
session, or the state information before you restart the session. PWXPC cannot guarantee recovery if you
make any of these changes.

When the PowerCenter Integration Service resumes a session, it restores the session state of operation,
including the state of each source, target, and transformation. PWXPC, in conjunction with the PowerCenter
Integration Service, determines how much of the source data it needs to reprocess.

PWXPC Restart and Recovery Operation 177

 Related Topics:
• “Recovering from CDC Session Failures” on page 184

Configuring CDC Sessions

Control PWXPC restart processing with session options and attributes. Ensure that all of the appropriate
session configuration is done.

The following table summarizes the options and attributes with defaults that must be changed for CDC
sessions:

Option / Attribute Location Recommended Value and Explanation

Commit Type Properties Tab Source.

The default is Target. The PowerCenter Integration Service
automatically overrides it to Source. You cannot change Commit
On End Of File unless you change Commit Type to Source.

Commit On End Of Properties Tab Disabled.

File The default is enabled. The PowerCenter Integration Service
performs a commit when the session ends. This commit occurs
after PWXPC commits the restart tokens. Duplicate data can
occur on restart if this default is enabled.

Recovery Strategy Properties Tab Resume from last checkpoint.

The default value is Fail task and continue workflow. PWXPC
requires Resume from last checkpoint.

Target Load Type Mapping Tab, Normal.

Targets Node To use recovery, you must select Normal instead of Bulk for
Target Load Type. Informatica recommends enabling recovery
for CDC sessions.

Stop on errors Config Object Tab 1. Set this value to 1 to ensure target data and restart token
integrity is maintained.
The default value is 0. Errors encountered by the PowerCenter
Integration Service while writing to the target database or target
files are not counted by default. Writer threshold errors can
include key constraint violations, loading nulls into a not null
field, and database trigger responses. PWXPC is unaware that
the writer has not applied the data to the target and so advances
the restart tokens.

Application Name Application Unique name.

Connection The default is the first 20 characters of the WorkFlow Name.The
default may not result in a unique name. Specify a unique name.

178 Chapter 6: Restart and Recovery

 Option / Attribute Location Recommended Value and Explanation

Restart Token File Application Unique name.

Name Connection The default is either:
- the WorkFlow Name if no Application Name is specified
- the Application Name
The default may not result in a unique name. Specify a unique
name.

Number of Runs to Application Value greater than 0.

Keep RestartToken Connection The default is 0. PWXPC keeps only one backup copy of the
File restart token initialization and termination files. Specify a value
greater than 0 so a history is available for recovery purposes.

Related Topics:
• “Restart Token File” on page 166
• “Application Names” on page 179
• “Enabling Session Recovery for a CDC Session” on page 177

Application Names
PowerExchange, with ODBC connections, stores the restart tokens in the PowerExchange CDEP file on the
extraction platform. PWXPC stores the restart tokens in:

• The state file on the PowerCenter Integration Service platform, for nonrelational sources
• The state table in the target database, for relational sources
You must specify a unique application name for CDC sessions to ensure proper restart processing.

Warning: Application names must not be shared with other CDC sessions. You may compromise data
integrity or experience data loss if you use the same application name for multiple, different CDC sessions.

Generating Current Restart Tokens for CDC Sessions

Use one of the following methods to generate current restart tokens for CDC data sources:

• The special override statement with CURRENT_RESTART in the PWXPC restart token file. PWXPC
and PowerExchange generate restart tokens that represent the current end of the change stream at the
time the PowerCenter session runs.
• The DTLUAPPL utility with the GENERATE RSTTKN parameters and a valid capture registration.
DTLUAPPL generates restart tokens that represent the current end of the change stream at the time the
utility runs.
• PowerExchange Navigator database row test function. PowerExchange generates restart tokens with
the location and source type you specify. These restart tokens represent the current end of the change
stream at the time at which you perform the database row test.

PWXPC Restart and Recovery Operation 179

 Related Topics:
• “Configuring the Restart Token File” on page 173

DTLUAPPL with CDC Sessions

You can use the PowerExchange DTLUAPPL utility to generate current restart tokens for CDC data sources.
You can then update the restart token file with the generated restart tokens to establish the starting extraction
point in the change stream for a CDC session.

Example of Generating Restart Tokens with DTLUAPPL

The registration on the ADD RSTTKN statement must exist. This registration must be associated with the
instance name specified on the MOD APPL statement. You can use either ADD APPL or MOD APPL for a
new application name. DTLUAPPL creates the application name if it does not exist when MOD APPL is used.

Tip: Select an application name and use it only with DTLUAPPL when generating restart tokens.

The following example generates restart tokens for source registration DB2DEMO1 with an application name
of tokens and then prints those restart tokens:
MOD APPL tokens DSN7 RSTTKN GENERATE
ADD RSTTKN db2demo1
END APPL tokens
PRINT APPL tokens

The DTLUAPPL prints the generated tokens because the PRINT APPL statement is specified:
Application name=<tokens> Rsttkn=<1> Ainseq=<0> Preconfig=<N>
FirstTkn =<>
LastTkn =<>
CurrentTkn=<>
Registration name=<db2demo1.1> tag=<DB2DSN7db2demo11>
Sequence=<000007248B9600000000000007248B9600000000>
Restart =<D2D1D4D34040000007248B0E00000000>

DTLUAPPL does not generate the complete restart1_token value which is shown in the SEQUENCE token.
You must be add the trailing four bytes (eight digits) of zeros manually when you update the restart token file.
DTLUAPPL does generate the complete restart_token2 value in the Restart= token. You can copy this value
to the restart token file.

Tip: You can use same restart tokens for multiple source tables in the restart token file to start extracting
changes from the same point in the change stream. You only need to run DTLUAPPL multiple times if you
want to start extracting changes from different locations in the change stream for different sources.

By using the tokens in this example, the restart token file looks as follows:
D1DSN7.db2demo1=000007248B9600000000000007248B960000000000000000
D1DSN7.db2demo1=D2D1D4D34040000007248B0E00000000

Starting CDC Sessions

Use the Workflow Manager, Workflow Monitor, or pmcmd to run a workflow or task. You can choose to run
the entire workflow, part of a workflow, or a task in the workflow. You can choose to warm start, recover, or
cold start workflows and tasks.

180 Chapter 6: Restart and Recovery

Cold Start Processing
Cold start workflows and tasks by selecting the cold start command in Workflow Manager, Workflow Monitor,
or pmcmd. When you request cold start for CDC sessions, PWXPC writes the following message in the
session log:
PWXPC_12091 [INFO] [CDCRestart] Cold start requested

PWXPC performs the following tasks for cold start:

1. Reads the restart tokens from the restart token file only.
2. Commits the restart tokens to the state tables and files, and then writes message PWXPC_12104 to the
session log.
3. Continues processing and committing the change data and restart tokens, until the session ends or is
stopped.

Related Topics:
• “Determining the Restart Point” on page 167

Warm Start Processing

Warm start workflows and tasks by selecting the start or restart command in Workflow Manager, Workflow
Monitor, or pmcmd. When you request warm start, PWXPC writes the following message to the session log:
PWXPC_12092 [INFO] [CDCRestart] Warm start requested. Targets will be resynchronized
automatically if required

PWXPC automatically performs recovery when you warm start a workflow or task. You do not need to
recover workflows and tasks before you restart them.

PWXPC performs the following tasks for warm start:

1. Reconciles the restart tokens from the restart token file and from the recovery state tables and files.
2. For heterogeneous targets, PWXPC queries the PowerCenter Integration Service about the commit
levels of all targets. If all targets in the session are at the same commit level, PWXPC skips recovery
processing.
3. If recovery is required for heterogeneous targets, PWXPC re-reads the data for the last UOW committed
to higher-level targets and flushes it to those targets with the lower commit level. The PowerCenter
Integration Service commits any flushed data and restart tokens to any relational targets and updates
any nonrelational files.
Note: PWXPC does not read the restart token file if recovery is required.
4. If recovery is not required and the reconciled restart tokens differ from those in the state tables and files,
PWXPC commits the reconciled restart tokens and issues message PWXPC_12104.
5. Continues processing and committing data and restart tokens until the session ends or is stopped.

Related Topics:
• “Determining the Restart Point” on page 167

Recovery Processing
Recover workflows and tasks by selecting the recover command in Workflow Manager, Workflow Monitor, or
pmcmd. When you request recovery, PWXPC writes the following message to the session log:
PWXPC_12093 [INFO] [CDCRestart] Recovery run requested. Targets will be resynchronized
if required and processing will terminate

PWXPC Restart and Recovery Operation 181

 Select Recover to populate the restart token file with the restart tokens for all sources in the CDC session so
that you can cold start. You can also use recovery to ensure the targets and restart tokens are in a consistent
state.

PWXPC automatically performs recovery when you warm start a workflow or task. You do not need to
recover workflows and tasks before you restart them. PWXPC performs the following tasks for recovery:

1. Reads the restart tokens from the recovery state tables and file.
2. Creates the initialization restart token file with the reconciled restart tokens.
3. For heterogeneous targets, PWXPC queries the PowerCenter Integration Service about the commit
levels of all targets. If all targets in the session are at the same commit level, PWXPC skips recovery
processing.
4. If recovery is required for heterogeneous targets, PWXPC re-reads the data for the last UOW committed
to higher-level targets and flushes it to those targets with the lower commit level. The PowerCenter
Integration Service commits any flushed data and restart tokens to any relational targets and updates
any nonrelational files.
Note: PWXPC does not read the restart token file if recovery is required.
5. Updates the restart token file with the final restart tokens, creates the termination restart token file, and
ends.
You can now warm start or cold start the workflow or task to process change data from the point of
interruption.

Stopping CDC Sessions

Stop CDC sessions through PowerCenter by issuing the stop command from Workflow Monitor or pmcmd.
You can also stop CDC sessions through PowerExchange by issuing the STOPTASK command or by running
the DTLUTSK utility. You can also abnormally terminate a workflow or task with the abort command in
Workflow Monitor or pmcmd.

When you stop a workflow or task gracefully by issuing the stop command through PowerCenter or
PowerExchange, the following actions occur:

1. If you issue the PowerCenter stop command, the PowerCenter Integration Service requests PWXPC to
stop.
If you issue the PowerExchange stop command, PowerExchange sends an end of file to PWXPC.
2. PWXPC performs end of file processing to flush the remaining uncommitted complete units of work and
their associated restart tokens to the targets. PWXPC then write messages PWXPC_12101 and
PWXPC_12068 to the session log.
3. The PowerCenter Integration Service processes all of data in the pipeline and writes it to the targets.
4. The PowerCenter Integration Service sends an acknowledgement to PWXPC indicating that the targets
have been updated.
5. PWXPC writes the termination restart token file, and then writes message PWXPC_12075 to the session
log. After this, the PWXPC reader shuts down.
6. The PowerCenter Integration Service ends the session successfully.
You can also stop CDC sessions gracefully if you use:

• Idle Time=0 in the PWX CDC Real Time connection. Idle Time=0 instructs PowerExchange to stop
processing at end of log.
• PWX CDC Change connection to extract changes from condense files. When you use PowerExchange
batch extraction mode for condense files, the extraction automatically ends when all condensed data is
read.

182 Chapter 6: Restart and Recovery

 Related Topics:
• “Configuring Idle Time” on page 119

Changing CDC Sessions and Restarting from the Point of

Interruption
You can add new sources and targets to an existing CDC sessions. However, you must cold start the session
after you do this, which means you need to ensure you have the proper restart information for the original
sources in the session.

To change a CDC session and restart it from the point of interruption:

1. Gracefully stop the workflow.

2. After the workflow stops successfully, issue the Recover command for the CDC session.
When you recover tasks, PWXPC writes the ending restart tokens for all sources in the session into the
restart token file specified on the CDC connection.
3. Change the session or workflow as desired.
4. Ensure that the restart token file in the source CDC connection points to the same restart token file
updated in the recovery session.
5. Optionally, add statements providing restart tokens to the restart token file for any new sources.
6. If you remove any sources from the session, update the restart token file to remove their restart tokens.
7. Cold start the CDC session.

Related Topics:
• “Stopping CDC Sessions” on page 182

Adding a Source to a Session - Example

In this example, a new source (RRTB_SRC_004) is added to an existing CDC session containing three
sources. The restart points for the existing sources are maintained. DTLUAPPL is used to generate a restart
token representing the current end-of-log in the change stream.

To add a new source to an existing session - example:

1. Stop the workflow by issuing the Stop command in Workflow Monitor.

2. After the workflow stops, use the Workflow Monitor and issue the Recover Task command from Workflow
Monitor to run a recovery session. PWXPC writes the following messages in the session log:
PWXPC_12060 [INFO] [CDCRestart]

===============================
Session restart information:
===============================
Extraction Map Name Restart Token 1 Restart Token 2 Source
d1dsn7.rrtb0002_RRTB_SRC_002 000000AD220F00000000000000AD220F0000000000000000 C1E4E2D34040000000AD0D9C00000000 GMD storage
d1dsn7.rrtb0001_RRTB_SRC_001 000000AD220F00000000000000AD220F0000000000000000 C1E4E2D34040000000AD0D9C00000000 GMD storage
d1dsn7.rrtb0003_RRTB_SRC_003 000000AD220F00000000000000AD220F0000000000000000 C1E4E2D34040000000AD0D9C00000000 GMD storage

PWXPC places the restart tokens in the restart token file specified in the CDC application connection.
3. Make any necessary changes to the mapping, session, and workflow to add the new source,
RRTB_SRC_004.
4. Run DTLUAPPL with RSTTKN GENERATE to generate restart tokens for the current end-of-log. Use the
following DTLUAPPL control cards to do this:
mod APPL dummy DSN7 rsttkn generate
mod rsttkn rrtb004

PWXPC Restart and Recovery Operation 183

 end appl dummy
print appl dummy
The PRINT command produces the following output:
Registration name=<rrtb004.1> tag=<DB2DSN7rrtb0041>
Sequence=<00000DBF240A0000000000000DBF240A00000000>
Restart =<C1E4E2D3404000000DBF238200000000>
Add eight zeroes to the end of the Sequence= value to create the restart token file value.
5. Update the restart token file to add the new source and its tokens. The updated file looks as follows:
<!-- existing sources
d1dsn7.rrtb0001_RRTB_SRC_001=000000AD220F00000000000000AD220F0000000000000000
d1dsn7.rrtb0001_RRTB_SRC_001=C1E4E2D34040000000AD0D9C00000000
d1dsn7.rrtb0002_RRTB_SRC_002=000000AD220F00000000000000AD220F0000000000000000
d1dsn7.rrtb0002_RRTB_SRC_002=C1E4E2D34040000000AD0D9C00000000
d1dsn7.rrtb0003_RRTB_SRC_003=000000AD220F00000000000000AD220F0000000000000000
d1dsn7.rrtb0003_RRTB_SRC_003=C1E4E2D34040000000AD0D9C00000000
<!-- new source
d1dsn7.rrtb0004_RRTB_SRC_004=00000DBF240A0000000000000DBF240A0000000000000000
d1dsn7.rrtb0004_RRTB_SRC_004=C1E4E2D3404000000DBF238200000000
6. Cold start the session.
PWXPC passes these restart tokens to PowerExchange to recommence extracting changes from the
change stream.
Note: This restart point is earlier than the one just generated for the new source. The new source does
not receive any changes until the first change following its restart point is encountered.

Related Topics:
• “Configuring the Restart Token File” on page 173
• “DTLUAPPL with CDC Sessions” on page 180

Recovering from CDC Session Failures

If you run a session with a resume recovery strategy and the session fails, do not edit the state information or
the mapping for the session before you restart the session.

CDC sessions fail for a number of reasons, such as:

• Permanent errors such as source or target data errors

• Transitory errors such as infrastructure problems, server crashes, and network availability issues
If the session fails because of transitory errors, restart the session after the source of the transitory error is
corrected. PWXPC automatically recovers warm started sessions, if required although you can also run a
recovery session.

Note: You cannot override the restart point if recovery processing is required. PWXPC does not read the
restart token file if you warm start and recovery is required or if you run a recovery session.

CDC sessions also fail because of permanent errors, such as SQL failures or other database errors. You
must correct permanent errors before restarting the CDC session. With some failures, you can correct the
error and then restart the CDC session. In other cases, you need to re-materialize the target table from the
source table before you recommence applying changes to it. If you re-materialize the target table, you need
to provide restart tokens matching the new restart point in the change stream and the cold start the CDC
session.

184 Chapter 6: Restart and Recovery

 Related Topics:
• “Recovery Processing” on page 181

Example: Session Recovery

In this example, a CDC session with relational targets is aborted in the Workflow Monitor. The Restart Task
command is issued from the Workflow Monitor for the CDC session. PWXPC issues the following message:
PWXPC_12092 [INFO] [CDCRestart] Warm start requested. Targets will be resynchronized
automatically if required

PWXPC automatically recovers when the session is warm started. PWXPC issues the following messages
displaying the restart tokens found for the session and its sources:
PWXPC_12060 [INFO] [CDCRestart]

===============================
Session restart information:
===============================
Extraction Map Name Restart Token 1 Restart Token 2 Source
d1dsn8.rrtb0004_RRTB_SRC_004 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0009_RRTB_SRC_009 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0005_RRTB_SRC_005 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0006_RRTB_SRC_006 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0008_RRTB_SRC_008 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0003_RRTB_SRC_003 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0002_RRTB_SRC_002 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0001_RRTB_SRC_001 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage
d1dsn8.rrtb0007_RRTB_SRC_007 00000FCA65840000000000000D2E004A00000000FFFFFFFF C1E4E2D3404000000D21B1A500000000 GMD storage

PWXPC issues the PWXPC_12069 message when it detects that recovery is required. The 12069 message
usually includes the begin-UOW (from) and end-UOW (to) restart tokens for the oldest uncommitted UOW
that PWXPC re-reads during recovery. PWXPC stores end-UOW restart tokens in the state table and file
unless sub-packet commit is used.
PWXPC_12069 [INFO] [CDCRestart] Running in recovery mode. Reader will resend the the oldest uncommitted UOW to resync targets:
from: Restart 1 [00000FCA65840000000000000D2E004A00000000FFFFFFFF] : Restart 2 [C1E4E2D3404000000D21B1A500000000]
to: Restart 1 [00000FCA65840000000000000D300D8000000000FFFFFFFF] : Restart 2 [C1E4E2D3404000000D21B1A500000000].

The from restart tokens are the same as those displayed in the PWXPC_12060 messages for all sources.
This restart token represents the start point in the change stream for the oldest uncommitted UOW. The to
restart tokens represent the end of the oldest uncommitted UOW. Since the application connection for this
session specifies sub-packet commit, the Restart 2 value is the begin-UOW value in both cases. The Restart
1 values represent the start and end change records in the Restart 2 UOW.

PWXPC rereads the changes between the two restart token values in the 12069 message and issues a
commit for the data and the restart tokens. The PowerCenter Integration Service writes the data to the target
tables and the restart tokens to the state table.

PWXPC and the PowerCenter Integration Service continue to read and write data and restart tokens until the
session ends or is stopped.

Related Topics:
• “Configuring Maximum Rows per Commit” on page 124

Managing Session Log and Restart Token File History

When you run CDC sessions, session log and restart token file history is often useful for recovery situations.
You can use these historical files to find previous restart points when data needs to be re-extracted.

By default, the PowerCenter Integration Service keeps a single session log for each workflow or session.
Change the Save session log for these runs option in either the session or Session Configuration
properties to specify the number or runs to keep the session logs. You can also change the
Save session log by option to keep session logs by timestamp instead of number of runs.

PWXPC Restart and Recovery Operation 185

 PWXPC keeps a single pair of initialization and termination restart token files. In the source CDC connection,
specify a value other than 0 or 1in the Number of Runs to Keep RestartToken File option to retain more
than a single pair of files.

You can determine starting and ending restart points for each extraction by using historical copies of the
restart token file. You need historical copies of the session logs to re-extract changes at any point in between
the session start and end.

When PWXPC issues a real-time flush to commit data to the targets, it issues message PWXPC_10081. This
message contains the restart tokens at that point in time:
PWXPC_10081 [INFO] [CDCDispatcher] raising real-time flush with restart tokens
[<restart1_token>], [<restart2_token>] <because UOW Count [<n>] is reached.> | <because
Real-time Flush Latency [<n>] occurred.>

To restart an extraction from a specific commit point, use the restart tokens in the appropriate 10081
message to populate the restart token file and cold start the CDC session. PWXPC passes the restart token
file values to PowerExchange to extract the data from that point forward.

186 Chapter 6: Restart and Recovery

CHAPTER 7

Flexible Target Key

Transformation
This chapter includes the following topics:

• Flexible Target Key Transformation Overview, 187

• Configuring PowerExchange Extraction Maps, 188
• Configuring Mappings with Flexible Target Key Transformations, 189
• Configuring Flexible Target Key Transformations with Multiple Sources, 190

Flexible Target Key Transformation Overview

The Flexible Target Key Custom transformation has been developed to address the case where a data target
has its Primary Key mapped to a column in a PowerExchange Change Data Capture source, which may be
subject to update.

Prior to the implementation of the Flexible Target Key facility, the only way to identify these changes to a key
field was to pass the full record before and after image in the change stream. For sessions involving large
amounts of data this overhead could be significant.

For those wishing to keep network traffic to a minimum, PowerExchange has been enhanced to allow
indicators to be added to specific fields. These indicators identify instances where that field has changed, and
comprise both a “before image” field and a “change indicator” for any given field. From this information a
delete/insert pair can then be generated for the target.

187
 The following figure shows the process:

COL1 is updated from an A to a C. The PowerExchange Listener sends a before-image column value of A
and an after-image column value of C to PowerCenter. Because COL1 is marked as the primary key for the
target table, the Flexible Target Key transformation processes the update as a delete of row A and an insert
of row C.

Configuring PowerExchange Extraction Maps

To ensure enough information is passed through to PowerCenter for a column, or columns, intended for use
as a target primary key, a “change indicator” and “before image” field must be associated with the
PowerExchange extraction map.

To configure a PowerExchange extraction map:

1. To add the before image and change indicator, right-click the extraction group.
2. Open the extraction group.
3. Open the extraction map. Either right-click the map and select Open, or double-click the extraction map.
4. Right-click the column that requires the before image and change indicator to be set.
5. Select Amend Control Indicator/Before Image Extensions.
6. To set the change indicator, select and add the required column by double-clicking or highlighting it and
then clicking Add.
7. When all the required change indicators have been set up, click the Before Images tab.
8. Repeat the process for columns that require the before image to be included with the change capture
data.
Note: The BI and CI column names can be changed on these screens by single-clicking and editing. The
following sections in this chapter will refer to the default names.

188 Chapter 7: Flexible Target Key Transformation

Configuring Mappings with Flexible Target Key
Transformations
The PowerExchange Change Data Capture (CDC) source will now need to be imported. This must be done
through the Import from PowerExchange option under Source Analyzer in the PowerCenter Designer.

After the CDC source has been imported, the target definitions will be required. Import them in the Target
Designer in the PowerCenter Designer.

The following figure shows an example structure of a source CDC table that works with a DB2 data capture
source and a DB2 target table:

Note: The columns prefixed DTL_CI and DTL_BI are the change indicator and before image, respectively.
These columns are necessary for this flexible transformation. The DTL__CAPX columns are PowerExchange
capture columns.

The following figure shows the structure of the target DB2 table:

The source table has a primary key of CUSTOMER_ID, and the PARTNER_CUST_ID is an attribute of the
table. The transformation will load the data into the target table where the PARTNER_CUST_ID is the
primary key, and the CUSTOMER_ID is an attribute of that table.

Configuring Mappings with Flexible Target Key Transformations 189

 In this case, if the PARTNER_CUST_ID changes in the source table, a delete and insert will be required for
the target.

Drag the PowerExchange CDC source, and the target into the mapping.

Note: Any of the control information including BI and CI columns could, if required, be defined in the target
definition.

To configure a mapping with a Flexible Target Key transformations:

1. From the Transformation option, select Create.

2. Select a transformation type of Flexible Target Key and click Create.
The transformation is inserted into the mapping.
3. Link into the transformation the required columns from the source for the target.
Also, the transformation will require the BI and CI indicators that were assigned earlier in
PowerExchange. A final column DTL__CAPXACTION should also be added to the transformation.
The BI and CI columns need to be linked to the Flexible transformation.
4. Right-click the transformation you have just created. Select Edit, and then select the Source Column
Map tab.
5. Add the columns to which the relevant before images and change indicators are assigned.
Multiple columns can be added here when the target primary key demands.
6. Link the transformation to the target.
Only link the required data columns from the transformation to the target. No links will exist from the
transformation to the target for the CI and BI columns, nor for the DTL__CAPXACTION column unless
required in the target.
7. Create the workflow from the mapping as normal.

Related Topics:
• “Working with Extraction Map Definitions” on page 44
• “Configuring PowerExchange Extraction Maps” on page 188
• “Source and Target Definitions” on page 28

Configuring Flexible Target Key Transformations with

Multiple Sources
The following example shows how to process changes captured from a VSAM file with two different record
types through a single Flexible Target Key transformation and then apply the changes to two DB2 target
tables.

This example assumes that the data map for the VSAM data set has been modified to assign the relevant
record id value test for both record types. In this example, a data map has been created for records with a
REC_TYPE of “A” and for REC_TYPE of “B”. These data maps are called ksdss1.ksdsm1 and
ksdss2.ksdsm2, respectively. These data maps have been registered for change data capture.

1. Assign the BI and CI PowerExchange fields to the extraction map.

2. Import the two extraction maps as data sources into the PowerCenter Designer by using the Import from
PowerExchange dialog box.

190 Chapter 7: Flexible Target Key Transformation

 Import the relevant data targets. In this example, the two input capture streams will be written to two
separate DB2 tables.
The following figure shows the mapping:

The Flexible Target Key transformation contains the columns for both of the PowerExchange sources,
and those sources each link to their own target tables. The Flexible Target Key transaction includes
input and output groups. An input and output group is added when the transformation is created.
Subsequent input/output groups are required for each source. These are added when you “Edit” the
transformation and select the Ports tab and then use the Create Input Group and Output Group buttons.
3. Before validating the mapping, you must assign the before image and change indicators to the ports.

Related Topics:
• “Configuring PowerExchange Extraction Maps” on page 188

Configuring Flexible Target Key Transformations with Multiple Sources 191

Part III: PowerExchange ODBC
This part contains the following chapters:

• Installing PowerExchange ODBC, 193

• Working with Mappings for PowerExchange ODBC, 196
• Configuring Connections for PowerExchange ODBC, 200
• Working with Sessions for PowerExchange ODBC, 203
• PowerExchange Restart and Recovery, 207

192
CHAPTER 8

Installing PowerExchange ODBC

This chapter includes the following topics:

• Installing PowerExchange ODBC Overview, 193

• Modifying the PowerExchange Configuration Files, 193
• Creating ODBC Data Sources, 194

Installing PowerExchange ODBC Overview

Before you use PowerExchange ODBC connections with PowerCenter, install and configure PowerExchange
on the PowerCenter Client and PowerCenter Integration Service machines. Also, install the PowerExchange
ODBC drivers on Windows.

After you install and configure PowerExchange, you can create PowerExchange ODBC data sources.

Note: When connecting to PowerExchange from PowerCenter, Informatica recommends that you use
PWXPC instead of PowerExchange ODBC. PWXPC has additional functionality as well as improved
performance and superior CDC recovery and restart.

For more information about PowerExchange ODBC, see the PowerExchange Reference Manual.

Modifying the PowerExchange Configuration Files

You must define nodes for the PowerExchange Listeners in the PowerExchange configuration file,
dbmover.cfg, on the PowerCenter Integration Service and Client machines. The NODE statements specify
the information that PowerExchange uses to communicate with Listeners that access the source or target
databases or files.

Note: You can also use PowerExchange local mode if the data resides on the same machine as the
PowerCenter Integration Service. In local mode, a PowerExchange Listener is not required. If you use local
mode, specify local in the Location property or the LOCATION parameter for the ODBC data source. You do
not need to update the PowerExchange dbmover.cfg file.

Local mode is not available on 32-bit Windows systems.

193
Creating ODBC Data Sources
Create ODBC data sources on the PowerCenter Client and PowerCenter Integration Service machines.

On the PowerCenter Client machine, create ODBC data sources so that PowerCenter can import the
metadata for PowerExchange sources and targets and you can preview data from PowerCenter.

On the machine that runs the PowerCenter Integration Service, create ODBC data sources so that
PowerCenter workflows can connect to the PowerExchange Listener and extract data for PowerExchange
sources or targets.

Creating an ODBC Data Source on Windows

Create a PowerExchange ODBC data source on Windows for the PowerExchange Client and for the
PowerExchange Integration Service if it runs on Windows. Use the ODBC Data Source Administrator.

1. To start the ODBC Data Source Administrator, perform one of the following actions:
• If you plan to create a 32-bit data source on a Windows 32-bit system or a 64-bit data source on a
Windows 64-bit system, open the Control Panel and click Administrative Tools. Then double-click
Data Sources (ODBC).
• If you plan to create a 32-bit data source on a Windows 64-bit system, enter the following command
at the command prompt:
%windir%\SysWOW64\odbcad32.exe
Enter this command if you installed the PowerCenter Client, which is a 32-bit application, on a
Windows 64-bit system.
The ODBC Data Source Administrator dialog box appears.
2. On the System DSN tab, click Add.
3. In the Create New Data Source wizard, select the Informatica PowerExchange driver from the list of
available drivers and click Finish.
If this driver is not available, see the PowerExchange Installation and Upgrade Guide.
The PowerExchange Data Source dialog box appears.
4. On the PowerExchange Data Source tab, define the ODBC data source.
Enter values for all of the properties that are displayed for the selected data source type.
5. On the General tab, enter optional information that does not depend on the data source type.
6. On the appropriate Properties tab, enter properties that are specific to the data source type or access
method.
The following table identifies the tabs on which to enter the information by data source or access type:

Data Source or Access Type Properties Tab

CAPX CAPX Properties tab

CAPXRT CAPXRT Properties tab

DB2 for i5/OS AS/400 Properties tab

DB2 for z/OS DB2/S390 Bulk Load Properties tab

7. Click OK.

194 Chapter 8: Installing PowerExchange ODBC

 The ODBC data source appears in the System Data Sources list on the System DSN tab.
8. Click OK.
For more information, see the PowerExchange Reference Manual.

Creating a PowerExchange ODBC Data Source on Linux or UNIX

If the PowerCenter Integration Service runs on a Linux or UNIX machine, create a PowerExchange ODBC
data source on that machine. To create the data source, update the appropriate odbc.ini file.

The data source entries in the odbc.ini file have the following format:
[data_source_name]
DRIVER=PowerExchange_full_install_path/libdtlodbc.sl
DESCRIPTION='descriptive_text'
LOCATION=data_source_node_from_dbmover.cfg
DBTYPE=access_method_for_file_or_database
… (other ODBC parameters as appropriate)

The data source name defined in the odbc.ini file is specified in the Connect String attribute of the ODBC
connection in PowerCenter. This connect string is used to load the PowerExchange ODBC driver and
connect to the specified location to extract or load data.

For more information, see the PowerExchange Reference Manual.

Related Topics:
• “Creating an ODBC Data Source on Windows” on page 194

Creating ODBC Data Sources 195

CHAPTER 9

Working with Mappings for

PowerExchange ODBC
This chapter includes the following topics:

• Working with Mappings for PowerExchange ODBC Overview, 196

• Working with Source and Target Definitions for PowerExchange Batch (ODBC), 197
• Working with Source Definitions for PowerExchange Change or Real-time (ODBC), 198

Working with Mappings for PowerExchange ODBC

Overview
A mapping is a set of source and target definitions linked by transformation objects that define the rules for
data transformation. Mappings represent the data flow between sources and targets.

Source and target definitions represent metadata for sources and targets. When you create a source
definition, its structure differs depending on the type of source it represents:

• Nonrelational sources require a multi-group source definition.

• Relational sources use a single-group source definition.
The source qualifier for a source definition also differs in structure depending on the type of source definition.

After you create a source or target definition, you can include it in a mapping to extract data from the source
or load data to the target. You can extract source data in batch, change, or real-time mode. You can use one
source definition and one mapping for all modes.

196
 Related Topics:
• “PowerExchange ODBC Drivers” on page 20

Working with Source and Target Definitions for

PowerExchange Batch (ODBC)
With PowerExchange ODBC, you can create source or target definitions to access VSAM and sequential data
sets (flat files), Adabas, DataCom, DB2, IDMS, or IMS databases. You create the source and target
definitions identically regardless of the database type. After you create a source or target definition, you can
edit it.

For nonrelational sources, the metadata imported is from the PowerExchange data map created through the
PowerExchange Navigator. In the case of DB2 for z/OS and DB2 for i5/OS, the metadata is obtained either
directly from the DB2 catalog or from a PowerExchange data map.

Importing a Source or Target Definition for PowerExchange ODBC

You can use ODBC to connect to a PowerExchange Listener to import a source or target definition. The data
map must be sent to the PowerExchange Listener before you can import the data map by using an ODBC
connection in Designer.

When you connect to the PowerExchange Listener, the Designer displays a list of schemas and tables.
Before you connect to the source or target, you can filter the list by schema name. Then select a table from
the list to create the source or target definition.

When you import a DB2 for z/OS or DB2 for i5/OS source definition through a PowerExchange Listener, the
Designer imports the primary keys.

1. In the Source Analyzer, click Sources > Import from Database to import a source definition. In the
Target Designer, click Targets > Import from Database to import a target definition.
The Import Tables dialog box appears.
2. In the Owner name field, you can enter filter criteria to narrow the list of objects that are retrieved.
When you import PowerExchange data maps for nonrelational sources or targets, ensure that the owner
name is the schema name from the data map and the ODBC data source has a DB Type of NRDB or
NRDB2. When the DB2 catalog is used for a DB2 for z/OS or DB2 for i5/OS source or target, ensure that
the owner is the owner of the DB2 tables and the DB Type is either DB2 or DB2400C.
Note: The only difference between NRDB and NRDB2 is whether a three- or two-tier naming format is
used for tables in the SQL statements for extracting or loading data:
schema.mapname.table for NRDB
schema.mapname_table for NRDB2
If the ODBC data source or target points to a PowerExchange Listener that is running with a
PowerExchange security setting of SECURITY=(1,x) or (2,x) in the PowerExchange DBMOVER
configuration file, you must provide a valid user ID and password. For a data source or target on i5/OS
or z/OS, you can enter a valid PowerExchange passphrase instead of a password. An i5/OS passphrase
can be from 9 to 31 characters in length. A z/OS passphrase can be from 9 to 79 characters in length. A
passphrase can contain uppercase and lowercase letters, numbers, spaces, and the following special
characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?

Working with Source and Target Definitions for PowerExchange Batch (ODBC) 197
 Note: The first character is an apostrophe.
3. Click Connect.
4. Select the table or tables that you want to import. To select multiple tables, use the Shift key or Ctrl key
or click Select All.
5. Click OK.
The source or target definition appears.
6. Click Repository > Save.

Working with Source Definitions for PowerExchange

Change or Real-time (ODBC)
With PowerExchange ODBC, you can create source definitions to access change data for VSAM data sets,
including CICS/VSAM, and ADABAS, Datacom, DB2, IDMS, IMS, MSSQL Server, and Oracle databases.

Importing a Change Data Capture Source Definition for

PowerExchange ODBC
You can import PowerExchange extraction maps to create source definitions for CDC sources for which you
use PowerExchange ODBC connections.

When you create capture registrations in the PowerExchange Navigator, corresponding extraction maps are
automatically generated. You can also manually create them.

1. In the Source Analyzer, click Sources > Import from Database.

The Import Tables dialog box appears.
2. In the Owner name field, you can enter filter criteria to narrow the list of objects that are retrieved.
In PowerExchange extraction maps, the owner is the first qualifier of the extraction map name. The
extraction map name has the following format:
DNinstance.regname_TABLENAME
Where:
• D is the default entry starts with D (any user-modified maps start with U)
• N is the database-specific identifier, such as 1 for DB2 for z/OS, 2 for IMS, 3 for DB2 for i5/OS, and
so on.
• instance is the instance name chosen for the source registration
• regname is the registration name chosen for the source registration.
For example, a DB2 for z/OS extraction map name could be d1dsn9.testdb2_KJM723TB.
When you import extraction maps to create CDC source definitions, the ODBC data source must have a
DB Type of CAPX or CAPXRT. These DB Type values instruct the PowerExchange Listener to select
extraction maps rather than data maps.
If the ODBC data source points to a PowerExchange Listener that is running with a PowerExchange
security setting of SECURITY=(1,x) or (2,x) in the DBMOVER configuration file, you must provide a valid
user name and password. For a data source on i5/Os or z/OS, you can enter a valid PowerExchange
passphrase instead of a password. An i5/OS passphrase can be from 9 to 31 characters in length. A

198 Chapter 9: Working with Mappings for PowerExchange ODBC

 z/OS passphrase can be from 9 to 79 characters in length. A passphrase can contain uppercase and
lowercase letters, numbers, spaces, and the following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
3. Click Connect.
4. Select the table or tables that you want to import.
To select multiple tables, use the Shift key or Ctrl key or click Select All.
5. Click OK.
The source definition appears.
6. Click Repository > Save.

Working with Source Definitions for PowerExchange Change or Real-time (ODBC) 199
CHAPTER 10

Configuring Connections for

PowerExchange ODBC
This chapter includes the following topics:

• Configuring Connections for PowerExchange ODBC Overview, 200

• Working with Connections for PowerExchange ODBC, 200

Configuring Connections for PowerExchange ODBC

Overview
Before PowerCenter can access a source or target in a session, you must configure connections in the
Workflow Manager. When you create or modify a session that reads from or writes to a database, you can
select only configured source and target databases. Connections are saved in the repository. For
PowerExchange ODBC, you configure relational database connections.

For more information about PowerExchange ODBC, see the PowerExchange Reference Manual.

Working with Connections for PowerExchange

ODBC
This section describes the connection types used for extracting and loading data as well as how to configure
connection attributes.

Connection Types for Extracting Source Data for PowerExchange

ODBC
The type of connection you configure depends on the extraction mode but not on the source type. However,
the source type does affect which ODBC data source is used in the Connect String when multiple Listeners
and platform-specific parameters exist.

200
 The following table describes the connection type to create based on extraction mode:

Extraction Mode Application Connection or Connection Type

Relational Database
Connection

Batch mode (nonrelational Relational Connection ODBC with NRDB and NRDB2 Data Source
data map)

Batch mode (relational) Relational Connection ODBC with Data Source of appropriate
database type (DB2, DB2400C, ADAUNLD, and
so on.)

Change mode Relational Database ODBC with CAPX Data Source

Connection

Real-time mode Relational Database ODBC with CAPXRT Data Source

Connection

Connection Types for Loading Target Data for PowerExchange

ODBC
All target connections are Relational Database Connection when loading data to PowerExchange targets.

The following table describes the target database types and their capability:

Database Type (Access Method) Insert Update Delete

Adabas Yes Yes Yes

Datacom No No No

DB2 (DB2) Yes Yes Yes

DB2 (DB2400C) No No No

IDMS No No No

IMS Yes Yes Yes

Sequential (NRDB/NRDB2) Yes No No

VSAM- KSDS (NRDB/NRDB2) Yes Yes Yes

VSAM-ESDS (NRDB/NRDB2) Yes No No

VSAM-RRDS (NRDB/NRDB2) Yes No No

Working with Connections for PowerExchange ODBC 201

 Configuring Attributes for Connections for PowerExchange ODBC
When you use ODBC connections, the majority of the attributes are specified in the ODBC data source or
target rather than in the PowerCenter connection.

In Workflow Manager, source and target ODBC connections are defined as relational connections with a
subtype of ODBC. For data sources, the ODBC connection definition process is the same for Batch, Change,
and Real-time processing. The only difference in these connections is the ODBC data source.

The following table shows the connection attributes in the Connection Object Definition dialog box for an
ODBC relational connection:

Connection Required or Description

Attribute Optional

Name Required Name for the relational database connection.

User Name Required User name for access to the data source or target.

Password Required A password for the specified user.

To connect to a source or target on i5/OS or z/OS, you can enter a valid
PowerExchange passphrase instead of a password. An i5/OS passphrase
can be from 9 to 31 characters in length. A z/OS passphrase can be from
9 to 79 characters in length when you use ODBC. A passphrase can
contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double quotation
marks (“), or currency symbols.
On z/OS, the allowable characters in the IBM IRRPHREX exit do not
affect the allowable characters in PowerExchange passphrases.
Note: On z/OS, a valid RACF passphrase can be up to 100 characters in
length. PowerExchange truncates passphrases longer than 100
characters when passing them to RACF for validation.
To use passphrases, ensure that the PowerExchange Listener runs with a
security setting of SECURITY=(1,N) or higher in the DBMOVER member.
For more information, see "SECURITY Statement" in the PowerExchange
Reference Manual.

Connect String Required Name of the ODBC data source or target.

Code Page Required Code page for the PowerCenter Integration Service to use to extract the
data from the data source.

Connection Optional Executes an SQL command with each database connection. Default is
Environment SQL disabled.

202 Chapter 10: Configuring Connections for PowerExchange ODBC

CHAPTER 11

Working with Sessions for

PowerExchange ODBC
This chapter includes the following topics:

• Working with Sessions for PowerExchange ODBC Overview, 203

• Extracting Data from PowerExchange in Batch Mode (ODBC), 204
• Extracting Data from PowerExchange in Change and Real-time Mode (ODBC), 205
• Configuring a Session to Load Data to PowerExchange Targets (ODBC), 206

Working with Sessions for PowerExchange ODBC

Overview
After you create mappings in the PowerCenter Designer, you can create a session and use the session in a
workflow to extract, transform, and load data. You create sessions and workflows in the Workflow Manager.

You can create a session in a workflow to extract data in batch, change, or real-time mode. You determine
how you want the PowerCenter Integration Service to extract the data when you configure the session. You
can also create a session to load data to a target.

After you configure the workflow, you must schedule the workflow.

Pipeline Partitioning for PowerExchange ODBC

Depending on your source or target database, you can increase the number of partitions in a pipeline to
improve session performance. Increasing the number of partitions allows the PowerCenter Integration
Service to create multiple connections to sources and targets and process partitions of data concurrently.
While processing data, the PowerCenter Integration Service may process data out of sequence due to the
varying rates at which the partitions process data.

When you create a session in a workflow, the Workflow Manager validates each pipeline in the mapping for
partitioning. You can specify multiple partitions in a pipeline if the PowerCenter Integration Service can
maintain data consistency when it processes the partitioned data.

Partition Types for Extracting Source Data for PowerExchange ODBC

If you are configuring a session to read source data, you can use partitioning if the session mode is batch.

203
 The following table describes the partition types for partitioning points when extracting source data in batch
mode:

Source Type Partitioning Point Partition Type

Nonrelational Application Multi-Group Source Qualifier Pass-through

Relational Source Qualifier Key range

Pass-through

Partition Types for Loading Target Data for PowerExchange ODBC

If you are configuring a session to write data to DB2 for z/OS and DB2 for i5/OS targets, you can specify all
partition types.

Extracting Data from PowerExchange in Batch Mode

(ODBC)
To extract data in batch mode, you must select the appropriate ODBC relational connection and configure
session properties. The connection you select and the session properties you configure differ depending on
the type of source data.

Configuring Properties for Nonrelational Batch Mode Sessions

(ODBC)
The relational connection for nonrelational sources should point to an ODBC data source that has a DBType
of either NRDB or NRDB2. You should use a connection with the same DBType as used when the source
was imported. That is, if you imported the source mapping by using NRDB, then use an NRDB ODBC
connection in the session connection as opposed to an NRDB2.

Note: The only difference between NRDB and NRDB2 is whether a three or two tier naming conventions is
used in the SQL statements to extract or load data. Nonrelational sources and targets mapped in
PowerExchange can be referred to as either NRDB or NRDB2. The format is as follows:

• NRDB: schema.datamapname.table
• NRDB2: schema.datamapname_table
To configure properties for a nonrelational batch mode session:

1. In the Task Developer, double-click a session with a nonrelational source to open the session properties.
2. Click the Sources view on the Mapping tab.
3. In the Reader field of the Readers settings, verify that Relational Reader is selected.
4. In the Connections Value field, select a nonrelational ODBC connection, that is, NRDB or NRDB2.
5. In the Properties settings, configure the Owner Name attribute.
At minimum, the schema name of the PowerExchange data map, or the Owner name that is displayed
when the source mapping is edited, must be specified to correctly construct the SQL statement during

204 Chapter 11: Working with Sessions for PowerExchange ODBC

 execution. PowerExchange SQL Escape Sequences can also be specified in this attribute field to
override specifications in the data map.
For a list of the available SQL escape sequences, see the PowerExchange Reference Manual.
The following example indicates how the Owner Name attribute can be configured to provide the schema
name for the source (seq) as well as an override for the physical file name in the data map
(dtldsn=new.dataset.name):
seq{dtldsn=new.dataset.name}
Note: SQL escape sequences and the Owner Name can be specified in any order in the attribute field.
6. Click OK.

Configuring Properties for Relational Batch Mode Sessions

(ODBC)
To extract data from a relational source in batch mode, you must select the appropriate ODBC relational
connection and configure session properties. The relational connection should point to an ODBC data source
that has the DBType of the source (DB2, DB2400C, DB2UDB, MSSQL or Oracle). You can then configure the
session properties as you would any other PowerExchange ODBC source.

To configure a properties for a relational batch mode session:

1. In the Task Developer, double-click a session with a relational source to open the session properties.
Click the Sources view on the Mapping tab.
2. In the Reader field of the Readers settings, verify that Relational Reader is selected.
3. In the Connections Value field, select the appropriate relational ODBC connection.
4. In the Properties settings, configure the Owner Name attribute.
At minimum, the owner name of the source table must be specified to correctly construct the SQL
statement during execution. However, if the owner name was specified in the source mapping, do not
specify it here.
5. Click OK.

Extracting Data from PowerExchange in Change and

Real-time Mode (ODBC)
To extract data in change and real-time mode, you must select an ODBC relational connection and configure
session properties. The connection you select for change mode should refer to an ODBC data source with a
DBType of CAPX whereas a real-time mode connection should refer to an ODBC data source with a DBType
of CAPXRT.

When you use ODBC to extract data from PowerExchange in either Change or Real-time mode, restart is
handled from within PowerExchange.

Extracting Data from PowerExchange in Change and Real-time Mode (ODBC) 205
 Configuring Properties for Change and Real-time Mode Sessions
(ODBC)
Use the following procedure to configure a session for Change or Real-time mode:

1. In the Task Developer, double-click a session with a relational source to open the session properties.
2. Click the Sources view on the Mapping tab.
3. In the Reader field of the Readers settings, verify that Relational Reader is selected.
4. In the Connections Value field, select a connection that points to an ODBC data source with the DBType
of CAPX (for Change) or CAPXRT (for Real-Time) and the appropriate Location value.
5. In the Properties settings, configure the Owner Name attribute.
At minimum, the schema name of the source extraction map must be specified to correctly construct the
SQL statement during execution. This name is the first qualifier of extraction map name shown in the
PowerExchange Navigator. It can also be determined by checking the Owner Name in the source
mapping. PowerExchange SQL Escape Sequences can also be specified in this attribute field to override
specifications in the data map. For a complete list of the SQL Escape Sequences available, see the
PowerExchange Reference Manual.
The following example indicates the manner in which the Owner Name attribute can be configured to
provide the schema name for the source (d6vsam) as well as an override for the application name
specified in the DBQual2/Application Name field in the ODBC data source (dtlapp=new_appname):
{dtlapp=new_appname}d6vsam
Note: SQL escape sequences and the Owner Name can be specified in any order in the attribute field.
6. Click OK.

Configuring a Session to Load Data to

PowerExchange Targets (ODBC)
To load data to a PowerExchange target, you must select a relational database connection. Then configure
properties for the session as you would any other relational target.

To configure a session to load data to a PowerExchange target:

1. In the Task Developer, double-click a session with a relational source to open the session properties.
2. Click the Targets view on the Mapping tab.
3. In the Writers settings, ensure that Relational Writer is selected.
4. In the Connections Value field, select a connection which points to an ODBC data source with
appropriate DBType and Location value.
5. In the Properties settings, configure the Table Name Prefix attribute.
At minimum, the schema name of the target table (if relational) or the PowerExchange data map (if
nonrelational) must be specified in order to correct construct the SQL statement during execution.
6. Click OK.

206 Chapter 11: Working with Sessions for PowerExchange ODBC

CHAPTER 12

PowerExchange Restart and

Recovery
This chapter includes the following topics:

• PowerExchange Restart and Recovery Overview, 207

• Restart and Recovery with PowerExchange ODBC, 207

PowerExchange Restart and Recovery Overview

Recovery and restart need to be considered when designing and configuring sessions and workflows by
using either PowerExchange Change or Real-Time. The considerations differ depending upon whether the
session uses PowerExchange Client for PowerCenter (PWXPC) or PowerExchange ODBC. This is primarily
due to the differences in where the restart information is maintained.

With PowerExchange ODBC, the restart information is controlled and maintained on the PowerExchange
Listener platform in the internal change information file (CDEP) by using the application name specified in the
connection.

Restart and Recovery with PowerExchange ODBC

PowerExchange ODBC uses the restart tokens associated with application names in the CDEP file to
determine the starting extraction point for a session. The CDEP file used is the one maintained by
PowerExchange on the platform targeted by the extraction. The CDEP file contains both begin and end
restart tokens for each extraction. This differs from PWXPC where only the end restart tokens are maintained
in the restart token file.

The restart tokens determine the baseline point from which to extract the change data for the source
database. Since both start and end restart tokens are available when you use ODBC, the extraction type
decides which ones will be used for restart. The extraction type is specified in the ODBC connection in the
CAPXTYPE parameter. This can be overridden at the session level by using the SQL Escape Sequence
DTLXTYPE.

If the extraction type specified is Since Last (SL) then the extraction will commence at the last successful
ending point. If it is Restart (RS) then it will commence at the starting point of either the last successful
extraction that was run or the one that is specified with the Reset Start Point in the extraction application
dialog box in the PowerExchange Navigator.

207
 Note: Unique application names should be used for each for each session to prevent any conflict in the
CDEP between multiple sessions.

The CDEP information for an extraction is only updated when a session ends successfully. Failed sessions
do not update the restart token information in the CDEP with the progress so far.

Warning: Failed sessions that are restarted will extract data from the last successful session (based on
CAPXTYPE specified). This means that there is the possibility that duplicated changes will be sent to the
target. In the event of a failure, you must either restore the targets to match the restart point prior to restarting
the session or design your PowerCenter session to handle the possibility of duplicate records.

Managing Restart Tokens with PowerExchange ODBC

PowerExchange on the source platform reads the restart tokens from the CDEP for each source included in
the session. The first time a session is run with a new application name, an entry for it will be automatically
created in the CDEP if it does not already exist.

If the application name already exists in the CDEP, PowerExchange will use the appropriate restart tokens to
commence extraction from the change stream. Which restart tokens are used depends upon the extraction
type (SL or RS) and whether new tokens have been supplied by DTLUAPPL since the last run. If new tokens
have been supplied by DTLUAPPL, then those tokens will be used regardless of the extraction type.

If there are sources included that have newer restart points, they will not be provided any records from the
change stream until their restart points have been reached. This prevents sources from being passed records
that they processed in the previous extraction run.

After a session completes successfully, PowerExchange updates the CDEP with the ending restart tokens.

CDEP Restart Tokens Example

The following example shows the contents of restart tokens in the CDEP for an application that contains a
restart token file for the DB2 sources DB2DEMO1, DB2DEMO2, and DB2DEMO3:
Application name=<odbc_db2demo13ac> Rsttkn=<3> Ainseq=<0> Preconfig=< >
FirstTkn =<D2D1D4D340400000003ED5D600000000>
LastTkn =<D2D1D4D340400000003ED5D600000000>
CurrentTkn=<>
Registration name=<db2demo1.1> tag=<DB2DSN7db2demo11>
Sequence=<000007337D8A00000000000007337D8A00000000>
Restart =<D2D1D4D34040000007337D0200000000>
Registration name=<db2demo2.1> tag=<DB2DSN7db2demo21>
Sequence=<000007337D8A00000000000007337D8A00000000>
Restart =<D2D1D4D34040000007337D0200000000>
Registration name=<db2demo3.1> tag=<DB2DSN7db2demo31>
Sequence=<000007337D8A00000000000007337D8A00000000>
Restart =<D2D1D4D34040000007337D0200000000>

The application name odbc_db2demo13ac was created by using DTLUAPPL with RSTTKN GENERATE. The
following example shows the control cards:
mod APPL odbc_db2demo13ac DSN7 RSTTKN GENERATE
add rsttkn db2demo1
add rsttkn db2demo2
add rsttkn db2demo3
end APPL odbc_db2demo13ac
print appl odbc_db2demo13ac

You can use either ADD APPL or MOD APPL for a new application name as with MOD APPL DTLUAPPL will
create the application name if it does not already exist.

After the new restart tokens are provided through DTLUAPPL, the session using this application name can be
started and it will use this restart tokens.

208 Chapter 12: PowerExchange Restart and Recovery

APPENDIX A

PowerExchange Interfaces for

PowerCenter Tips
This appendix includes the following topics:

• Organizing Sources by Map Type in the Designer, 209

• Filtering Source Data with PWXPC, 210
• DTLREXE to Submit MVS Batch Jobs, 212
• Empty Files, 213

Organizing Sources by Map Type in the Designer

Source metadata imported in Source Analyzer in the PowerCenter Designer with PWXPC is stored into sub-
folders based on source type and location. The user cannot choose the sub-folder name and the table names
within those sub-folders must be unique.

With the PWXPC Import from PowerExchange dialog box, these folder names have the following format:
<source_type_location>. The source_type value equates to the Source Type specified in the Import from
PowerExchange dialog box. The location value equates to the Location specified in the Import from
PowerExchange dialog box and is a user-specified name in the PowerExchange dbmover.cfg file.

For example, if the Location is MVS2 and the Source Type is IMS, then the folder will be called IMS_MVS2.
All IMS sources imported from that location will be stored into that folder. You can use the name you assign
for the location as a method of organizing your source metadata in Designer.

There are some nonrelational source types where the schema name is different but map name can be the
same for both PowerExchange data map and extraction map (CDC Datamap) created from that data map.

To illustrate, assume that the following information is used to create the PowerExchange data map for an IMS
database:

• IMS database name is IMS1T01. This name is used as the PowerExchange data map name.
• A segment in that database results in a table in the data map that is called IMSSEG1.
• The PowerExchange schema name that used when creating the data map is IMS.
The PowerExchange map name (in the NRDB2 form) is constructed as follows:

schema_name.datamap_name_table_name

209
 In this example, the PowerExchange data map name is IMS.IMS1T01_IMSSEG1. If you use the IMS
database name for the Capture Registration Name as well, the resulting extraction map name will be
d2reconid.IMS1T01_IMSSEG1.

The map name for both the data map and extraction map are the same. Since the schema name is not used
in the source definition table name, the names will be exactly the same when imported in Designer. As a
result, only one will be able to be imported from the same Location. So, there are two choices:

1. Use a different name for the capture registration so that a unique extraction map name is created.
2. Use a different Location name to import data maps and extraction maps.
If you want to use the same names for both the data map name and the capture registration name, then you
will need to use a different location name in the Import from PowerExchange to place these two source
definitions in different folders.

To help organize sources in Designer, use location names in the PowerExchange dbmover.cfg that indicate
the type of data you are retrieving. Creating multiple NODE= statements (location names) for the same
listener is perfectly acceptable. In this example, if you create a NODE called CDCMAPS in the dbmover.cfg
and use this as the Location when importing the IMS extraction map, it will be stored in a sub- folder called
IMS_CDCMAPS.

This strategy of separating extraction maps from regular relational and nonrelational source metadata can be
a useful way of organizing source metadata in Designer.

Filtering Source Data with PWXPC

You can use PWXPC to filter data at the source so PowerExchange only sends the filtered rows to the
PowerCenter session. There are many reasons why you might want to do this. Some examples of scenarios
in which this capability is useful are:

• For some source types, such as VSAM and IMS, you cannot limit change capture changes based on only
certain columns changing.
• You can, for certain source types such as DB2 for i5/OS, create capture registrations in PowerExchange
that register only the columns you select. However, if the RDBMS logs are used directly for extraction and
do not support selective column capture, PowerExchange extracts the row even if none of the columns of
interest have changed.
• The source type is one where either PowerExchange itself or the RDBMS will only capture changes based
on columns of interest. However, additional or all columns are registered for capture because other
extractions require them.
• You only want to extract columns with a specific value. For example, you want to read all of the columns
in a table for a specific customer.
You can use the Filter Override attribute in the Session Properties to filter records for sources that use PWX
Batch, CDC Change, and CDC Real Time connections. When you specify filter conditions in the Filter
Override attribute, PWXPC includes a WHERE clause with the filter conditions on the SELECT statement
that is passed to PowerExchange. You must use proper SQL syntax for these overrides to prevent SQL
failures. Additionally, PowerExchange supports a limited set of SQL syntax for nonrelational sources, which
includes bulk and change data.

For more information about NRDB SQL syntax that PowerExchange supports, see the PowerExchange
Reference Manual.

210 Appendix A: PowerExchange Interfaces for PowerCenter Tips

 In the filter condition, you can specify any column that exists in the source definition, including
PowerExchange-generated columns such as DTL__CAPX columns, change indicator columns (DTL__CI) and
before image columns (DTL__BI).

PWXPC supports two forms of filter condition syntax. For single record source definitions such as CDC
extraction maps, relational tables, or single record nonrelational data maps, specify a single filter condition
statement. You can specify a single filter condition or join numerous filter conditions by using the conditional
operands that PowerExchange supports for NRDB SQL statements. For example:
column1 is NULL and column2=’A’

For multi-record nonrelational source definitions, you can also use the following syntax:
group_name1=filter_condition;group_name2=filter_condition;...

Use the group_name form to specify filter conditions for one or more record types in a multi-record source
definition. To apply a filter condition to all records in a multi-record source definition, use a single filter
condition without group_name. You cannot combine single filter conditions with group_name conditions.

Filter Override Examples

The following examples show how to use filter overrides for CDC and multi-record source definitions.

Example 1: Filter Change Data with Change Indicator Columns

To filter the change data for this source table to include changes when the ACCOUNT column was changed,
code the following filter condition:
DTL__CI_ACCOUNT=’Y’

To filter change data to include changes where the ACCOUNT column was changed to 9999, code the
following filter condition:
DTL__CI_ACCOUNT=’Y’ and ACCOUNT=9999

Note: To use DTL__CI change indicator or DTL__BI before-image columns in filter conditions, you must alter
the extraction map in PowerExchange Navigator to define these columns.

Example 2: Filter Records in a Multi-Record VSAM Source Definition

In the following example, a multi-record VSAM source definition contains the following four records:

• V07A_RECORD_LAYOUT
• V07B_RECORD_LAYOUT
• V07C_RECORD_LAYOUT
• V07D_RECORD_LAYOUT.
Each record has unique field names, which means you must use the group_name syntax for filter conditions.
To filter data for the first two records, the filter condition uses the following group_name syntax:
V07A_RECORD_LAYOUT=V07A_RECORD_KEY=1;V07B_RECORD_LAYOUT=V07B_RECORD_KEY=2

No filtering is done on the other two records in the multi-record source definition.

PWXPC creates a SELECT statement for each record in the multi-record source definition. Because the filter
override uses the group_name syntax and only specifies two records, PWXPC includes the WHERE clause
on the SELECT statements for records V07A_RECORD_LAYOUT and V07B_RECORD_LAYOUT.

To use a single filter condition to filter all records in a multi-record source definition, the filter condition must
apply to all records. Otherwise, the session fails. For example, if all records in a multi-record source definition

Filtering Source Data with PWXPC 211

 contain a column called ADDRESS, you can code the following filter condition to select all records where
ADDRESS is not null:
ADDRESS is not NULL

PWXPC applies this filter condition to all records in the multi-record source definition by including a WHERE
clause with the condition on the SELECT statement for each record.

DTLREXE to Submit MVS Batch Jobs

PowerExchange provides a utility program called DTLREXE which can be used to cause remote execution of
a program on another platform running PowerExchange. One option of this utility is the SUBMIT option which
provides the ability to submit a MVS batch job remotely.

This utility can be invoked in PowerCenter workflows with PowerCenter Command Tasks. Command tasks
can either be stand-alone tasks or pre/post session commands. Which you choose depends upon what you
want to do. There are advantages in using a stand-alone command task for repeatable functions as it only
needs to be coded once and can then be used by numerous workflows. On the other hand, it is possible to
easily configure a session to fail if a pre-session command task fails.

Some examples of the type of tasks that can be performed with DTLREXE PROG=SUBMIT are:

• Truncate a database table prior to loading data into it in a session. This is useful for database types for
which PowerCenter does not support truncate, such as Adabas.
• Notify a MVS-based job scheduler that the workflow is starting or ending. Some job scheduler products
provide batch posting utilities and they can be utilized with DTLREXE to submit a batch job.
• Unload a database to a flat file so it can then be used in a session to load another database.
• Clean up DB2 bulk load files when the session completes successfully.
• Submit any type of MBS Batch JOB for which waiting for the completion and returning a set of messages
is required.
In the Edit - Command dialog box, you can enter a command. The following code is an example of a
DTLREXE PROG=SUBMIT command entered as a pre-session command:
dtlrexe prog=submit loc=mvs fn=\"dtlusr.jcl(db2load)\"
mode=(job,wait) output=dtlusr.output
result=c:\submit\output\output.txt
uid=user01 pwd=pass01

In this example, the DTLREXE command specifies mode=(job,wait) which means that the DTLREXE will wait
for the job to complete. This, in turn, will cause the session to wait until this pre-session command completes.
In the “Error Handling” section of the Config Object, you can specify how to handle errors for pre-session
commands in the On Pre-session command task error field.

Note: Ensure that the JOB submitted through DTLREXE includes the appropriate DTLNTS steps if WAIT
mode is requested. The PowerExchange RUNLIB, in member DTLREXE, contains sample JCL to be used
with DTLREXE that includes the required DTLNTS steps.

If you use a stand-alone command task to submit a batch JOB with DTLREXE, there are no session
configuration options to check for success or failure. If you want to test the status of the command task in the
following session, you will need to use one of the task-specific workflow variables available in the Workflow
Manager, that is, PrevTaskStatus or Status. These variables can be used in link conditions to test the status
of tasks in a workflow.

212 Appendix A: PowerExchange Interfaces for PowerCenter Tips

 The link condition is created by double-clicking on the link between the DTLREXE command task and the
s_bulk_db2demo123_db2demoabc session to which it is connected. This action invokes the Expression
Editor, enabling you to add the test to ensure that the DTLREXE command task succeeded.

Empty Files
PowerExchange can automatically create flat files, new generations of a Generation Data Group (GDG), or
sequential data sets. When a PowerCenter session writes the first record to a file that does not exist,
PowerExchange creates the file. Otherwise, if the session does not write any records to the file,
PowerExchange does not create it. Some applications require that a file is created, even if it is empty, each
time a session or workflow runs.

To create an empty flat file, generation of a GDG, or sequential data set, code a CREATEFILE command in
the Pre SQL or the Post SQL properties for any PowerExchange source or target. PowerCenter uses the
connection information from the source or target on which you code the command and passes the
CREATEFILE command to PowerExchange. PowerExchange processes the command and creates the file.

To create a file before the session runs, code the CREATEFILE command in the Pre SQL property. To create
a file after the session runs, code the CREATEFILE command in the Post SQL property.

The following rules and guidelines apply to the CREATEFILE command:

• For most sources or targets that use a PowerExchange connection, you can code CREATEFILE
commands in the Pre SQL or the Post SQL property. However, you cannot code commands for data map,
DB2 unload, and extraction map sources because these sources do not have Pre SQL or Post SQL
properties.
• You do not need to create a PowerExchange data map or PowerCenter source or target definition for the
file name that you code in the command.
• For MVS and i5/OS, you can specify allocation information for the file by coding allocation parameters on
the CREATEFILE command. Otherwise, the PowerExchange Listener that allocates the files uses
allocation values from the DBMOVER configuration file, or default values.
Note: Because PowerExchange creates files either before or after a session runs, it does not use any
allocation information specified in session properties for nonrelational sources or targets.
• If you specify the CREATEFILE command in the Pre SQL attribute, do not use writer partitioning for the
session. If you use writer partitioning, the session executes the CREATEFILE command in the first
partition and then fails when trying to execute it again in another partition.

Creating Empty Files in a PowerCenter Session

You can create an empty flat file, a new generation of a GDG, or a sequential data set in a PowerCenter
session.

1. In Workflow Manager, right-click the appropriate task in the Task Developer or the workflow in Workflow
Designer.
2. Click Tasks > Edit.
The Edit Tasks dialog box appears.
3. Click the Mappings tab.

Empty Files 213

 4. In the Pre SQL or Post SQL attribute under Session-Level Properties, enter the CREATEFILE
command. Use the following syntax:
<CMD>CREATEFILE FN=file_name parameter2 parameter3 ...
The following table describes the parameters that you can enter for the command:

Parameter Operating Description

System

BS z/OS Block size.

Valid values are 1 to 32760.

CLOSEDEALLOC z/OS Deallocate the data set at CLOSE rather than at step termination
(FREE=CLOSE).
Valid value is Y.

EPWD z/OS or i5/OS An encrypted password for the specified user ID in the UID
parameter. Alternatively if you are creating a file on z/OS, you can
enter an encrypted PowerExchange passphrase.
Do not encrypt a passphrase that contains invalid characters, such
single quotation marks ('), double quotation marks ("), or currency
symbols.
Do not also specify the PWD parameter.

FN All Required. File or data set name to be created. You must enclose file
names that include parentheses in double quotes (“).
To create a new generation of a GDG, code:
FN=”gdg_base_name(+1)”
Restriction: If you create a new generation, the GDG base name
must exist. The PowerExchange Listener on MVS that creates the
new generation must specify Y for the GDGLOCATE parameter in the
DBMOVER configuration file on MVS.

LRECL z/OS or i5/OS Logical record length.

Valid values are 1 through 32756

214 Appendix A: PowerExchange Interfaces for PowerCenter Tips

Parameter Operating Description
System

PWD z/OS or i5/OS A password for the specified user ID in the UID parameter.
If you are creating a file on i5/OS or z/OS, you can enter a valid
PowerExchange passphrase instead of a password. An i5/OS
passphrase can be from 9 to 31 characters in length. A z/OS
passphrase can up to 128 characters in length if you use a PWXPC
connection or up to 79 characters in length if you use an ODBC
connection. A passphrase can contain the following characters:
- Uppercase and lowercase letters
- The numbers 0 to 9
- Spaces
- The following special characters:
’ - ; # \ , . / ! % & * ( ) _ + { } : @ | < > ?
Note: The first character is an apostrophe.
Passphrases cannot include single quotation marks (‘), double
quotation marks (“), or currency symbols.
Note: A valid RACF passphrase can be up to 100 characters in
length. PowerExchange truncates passphrases longer than 100
characters when passing them to RACF for validation.
Do not also specify the EPWD parameter.

MODELDCB z/OS Model DSCB to be used for the file creation. Generally, this is only
required for GDG data sets that are not SMS-managed.

RECFM z/OS or i5/OS Record format.

RELEASE z/OS Release unused, allocated space on CLOSE.

Valid value is Y.

RELPOS z/OS Relative position of the data set on a tape volume, or data set
sequence number.
Valid values are 1 to 512.

SPACE z/OS Space allocation parameters. Use the following syntax:

SPACE=(U, P, S)
Where:
- U is unit type. Valid values are: T for tracks or C for cylinders.
- P is primary space. Valid values are 1 through 16777215 tracks or the
equivalent in cylinders.
- S is secondary space. Valid values are 1 through 16777215 tracks or
the equivalent in cylinders.
If you enter a primary or secondary space value greater than 65535
tracks, coordinate with the space management team for the target
system to verify that such a large allocation is supported.

UID z/OS or i5/OS User ID. This parameter is required if the PowerExchange Listener is
configured for security (SECURITY=1 or 2).
Also specify either PWD or EPWD, but not both.

Empty Files 215

 Parameter Operating Description
System

UNIT z/OS Unit type.

VOLSER z/OS Volume serial.

Use a space to separate parameters in the command.

5. Click OK.

Empty File - Example

The following example CREATEFILE statements show how to create a new generation of a GDG and a flat
file.

To create a new generation with a record length of 80 bytes and a block size of 8880 in a GDG data set
called MY.GDG, code the following command:
<CMD>CREATEFILE FN=”MY.GDG(+1)” LRECL=80 BS=8880 RECFM=PS

The PowerExchange Listener on MVS that creates the new generation uses its DBMOVER member to
determine the remaining DCB and space attributes for this file.

To create a new flat file called my.flat.file, code the following command:
<CMD>CREATEFILE fn=my.flat.file

216 Appendix A: PowerExchange Interfaces for PowerCenter Tips

APPENDIX B

Datatypes and Code Pages

This appendix includes the following topics:

• Datatypes and Code Pages Overview, 217

• PowerExchange Nonrelational Datatypes and Transformation Datatypes, 217
• Restrictions on Relational Datatypes, 220
• Reading and Writing Binary Data in PowerExchange Client for PowerCenter , 221
• Code Pages, 221

Datatypes and Code Pages Overview

PowerCenter uses the following datatypes when reading source data, transforming the data, and writing
target data:

• Native datatypes. Specific to the source and target databases or PowerExchange. Native datatypes
appear in source and target definitions.
• Transformation datatypes. Generic datatypes that appear in transformations. The PowerCenter
Integration Service uses the datatypes to move data across platforms.

PowerExchange Nonrelational Datatypes and

Transformation Datatypes
You import PowerExchange data maps to create nonrelational source definitions. The datatypes of the fields
in the source definition match the datatypes for the corresponding fields in the data map. These datatypes
map to PowerCenter transformation datatypes. The transformation datatypes are displayed in the Application
Multi-Group Source Qualifier for the nonrelational source and other transformations in a mapping.

Note: In PowerExchange, you can specify a precision of up to 31 for certain field datatypes, such as
PACKED. These fields are mapped to the Decimal transformation datatype in PowerCenter. If you enable
high precision in the session properties, PowerCenter supports a maximum precision of 28 for the Decimal
datatype. If you do not enable high precision, or if you specify a precision greater than 28, PowerCenter uses
the Double datatype with a precision of 16 and data rounding might occur.

217
 The following table shows the PowerExchange nonrelational datatypes and the corresponding transformation
datatypes:

PowerExchange Precision Transformation Range

Datatype Datatype

BIN 10 Binary 1 to 104,857,600 bytes

You can pass binary data from a source to a target,
but you cannot perform transformations on binary
data. PowerCenter does not support binary data for
COBOL or flat file sources.

CHAR 10 String 1 to 104,857,600 characters

Fixed-length or varying-length string.

DATE 10 Date/Time Jan 1, 0001 A.D. to Dec 31, 9999 A.D.

Combined date/time value, with precision to the
nanosecond.

DOUBLE 18 Double Precision of 15 digits

Double-precision floating-point numeric value.

FLOAT 7 Double Precision of 15 digits

Double-precision floating-point numeric value.

NUM8 3 Small Integer Precision of 5 and scale of 0

Integer value.

NUM8U 3 Small Integer Precision of 5 and scale of 0

Integer value.

NUM16 5 Small Integer Precision of 5 and scale of 0

Integer value.

NUM16U 5 Integer Precision of 10 and scale of 0

Integer value.

NUM32 10 Integer Precision of 10 and scale of 0

Integer value.

NUM32U 10 Double Precision of 15 digits

Double-precision floating-point numeric value.

NUM64 19 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

218 Appendix B: Datatypes and Code Pages

PowerExchange Precision Transformation Range
Datatype Datatype

NUM64U 19 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

NUMCHAR - String 1 to 104,857,600 characters

Fixed-length or varying-length string.

PACKED 15 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

TIME 5 Date/Time Jan 1, 0001 A.D. to Dec 31, 9999 A.D.

Combined date/time value, with precision to the
nanosecond.

TIMESTAMP 5 Date/Time Jan 1, 0001 A.D. to Dec 31, 9999 A.D.

Combined date/time value, with precision to the
nanosecond.

UPACKED 15 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

UZONED 15 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

VARBIN 10 Binary 1 to 104,857,600 bytes

You can pass binary data from a source to a target,
but you cannot perform transformations on binary
data. PowerCenter does not support binary data for
COBOL or flat file sources.

PowerExchange Nonrelational Datatypes and Transformation Datatypes 219

 PowerExchange Precision Transformation Range
Datatype Datatype

VARCHAR 10 String 1 to 104,857,600 characters

Fixed-length or varying-length string.

ZONED 15 Decimal Precision 1 to 28 digits, scale 0 to 28

Decimal value with declared precision and scale.
Scale must be less than or equal to precision. If you
pass a value with negative scale or declared precision
greater than 28, the PowerCenter Integration Service
converts it to a double.

Restrictions on Relational Datatypes

With certain exceptions, the PowerExchange Client for PowerCenter (PWXPC) supports the same datatypes
for DB2 for z/OS, DB2 for i5/OS, and DB2 for Linux, UNIX, and Windows that PowerCenter supports. PWXPC
also supports the same Oracle and Microsoft SQL Server datatypes that PowerCenter supports.

LOB Datatypes
PowerExchange does not support large objects (LOB) datatypes. PWXPC does not include LOB columns
when you import metadata for tables that include these datatypes.

Note: If you use ODBC to import metadata for tables that include LOB datatypes, you cannot use this
metadata with PowerExchange.

DB2 for z/OS TIMESTAMP Datatype

How PowerCenter handles DB2 for z/OS TIMESTAMP columns depends on the DB2 version.

For DB2 versions earlier than version 10, the DB2 TIMESTAMP datatype maps to the PowerCenter
transformation date/time datatype. The date/type datatype treats subsecond values as 32-bit integer values
of up to nine digits.

The following table summarizes how the source TIMESTAMP datatype maps to the transformation date/time
datatype for DB2 versions earlier than version 10:

DB2 Datatype Range Transformation Range

Datatype

TIMESTAMP 26 bytes date/time Jan 1, 0001 A.D. to Dec 31, 9999

Precision 26, scale 6 A.D.
Precision is to the Precision is to the nanosecond.
microsecond.

DB2 10 for z/OS introduced support for extended-precision TIMESTAMP columns, which can have
subsecond values of up to 12 digits. The PowerExchange Client for PowerCenter supports DB2 10 extended-
precision timestamps. Because the transformation date/time datatype does not support this level of extended
precision, the DB2 10 TIMESTAMP datatype maps to the transformation string datatype, except when scale

220 Appendix B: Datatypes and Code Pages

 is 6. To maintain compatibility with pre-version 10 TIMESTAMP columns, which always have a scale of 6, all
DB2 TIMESTAMP columns that have a scale of 6 map to the date/time datatype.

The following table summarizes how the DB2 10 extended-precision TIMESTAMP datatype maps to
transformation datatypes based on scale:

Scale Precision Transformation Datatype

6 26 date/time

0 19 string

1 to 5 or 7 to 12 20+scale string

When you write DB2 TIMESTAMP data to a DB2 target, the source and target must have the same scale.
Otherwise, unpredictable results can occur. If the DB2 source TIMESTAMP field has a scale greater than 9
and the corresponding DB2 target TIMESTAMP field has a scale of 6, the value is truncated before it is
written to the target.

When you write DB2 10 TIMESTAMP data to a nonrelational target, define the following session attributes in
the Edit Tasks dialog box to ensure that all TIMESTAMP fields have the same format:

• Set the Date Time Format String to YYYY-MM-DD HH24:MI:SS.

• Clear Pre 85 Timestamp Compatibility.

Reading and Writing Binary Data in PowerExchange

Client for PowerCenter
If you read binary data from any source, or if you write binary data to a DB2 for z/OS or DB2 for i5/OS target
with the PowerExchange Client for PowerCenter, the data cannot be larger than 128000 bytes. Otherwise,
the PowerCenter Integration Service truncates the data.

Code Pages
When you configure a connection object in the Workflow Manager, select a code page for the connection.

If you configure thePowerCenter Integration Service for code page validation, the source code page must be
a code page that is a subset of the target code page.

If you configure the PowerCenter Integration Service for relaxed code page validation this is not the case.
You can select any code page supported by PowerCenter for the target.

Related Topics:
• “Oracle CDC Application Connections” on page 105

Reading and Writing Binary Data in PowerExchange Client for PowerCenter 221
APPENDIX C

PowerExchange Interfaces for

PowerCenter Troubleshooting
This appendix includes the following topic:

• Troubleshooting for PowerExchange Interfaces for PowerCenter, 222

Troubleshooting for PowerExchange Interfaces for

PowerCenter
When I go into Designer, I get messages about failures to load DLLs.
This can happen when PowerExchange Client for PowerCenter plug-ins are installed but cannot be loaded
for various reasons like incorrect releases of PowerExchange installed or PATH problems. For more
information, see Informatica Knowledge Base Article # 15346.

I want to import a DB2 for i5/OS source definition, but need to determine the name of the DB2 database
on the i5/OS machine.
Use the DSPRDBDIRE command to see a list of databases on the i5/OS machine.

The session failed with an error stating that the PowerExchange message repository cannot be loaded.
You can receive this error on UNIX when there is no PWX_HOME environment variable set to the
PowerExchange installation directory. Set the PWX_HOME environment variable to the PowerExchange
installation directory.

I set the Idle Time session condition to -1. However, the session completed with the following message:
Idle Time limit is reached.
This can occur if EOF=Y is specified in the PowerExchange configuration file (dbmover.cfg)
CAPI_CONNECTION statement. When you set EOF=Y, PowerExchange returns an EOF (which stops the
session) when it reaches the end of the change stream as determined at the time the session starts reading
from it. As a result, the PowerCenter session completes instead of continuing to run.

This message can also occur if the connection with PowerExchange is stopped through the PowerExchange
STOPTASK command.

222
My session seems to be processing the pipelines serially.
You can configure the PowerCenter Integration Service to process master and detail pipelines sequentially.
As a result, it reads data from each source in change and real-time modes sequentially.

Clear the PMServer 6.X Joiner Source Order Compatibility option on the Compatibility and Database tab in
the Informatica Server Setup. When you rerun the session, the PowerCenter Integration Service will process
pipelines concurrently.

The session failed with a plug-in error:

MAPPING> SDKS_38007 Error occurred during [initializing] reader plug-in #30nnnn.

This is a generic message indicating the PWXPC encountered an error. Review the session log for other
messages indicating what the problem is. If there are no other error messages in the session log, check the
PowerExchange logs on both the PowerCenter Integration Service platform and the Listener platform.

I want to read all of the changes I have captured and have them be inserts into a staging area. How do I
do this?
When you use PowerExchange ODBC to read captured changes, INSERT is the default operation. If you
want to apply the changes to the target by using the same operation as done on the source (INSERT,
UPDATE, or DELETE), you need to explicitly include an Update Strategy transformation in the mapping to
make this happen by testing the DTL__CAPXACTION field. In the Update Strategy Expression field, you
would code:
DECODE(DTL__CAPXACTION,'I',DD_INSERT,'U',DD_UPDATE,'D',DD_DELETE,DD_REJECT)

When you use PWXPC, the DTL__CAPXACTION field is automatically acted upon when processing change
data. If you want to have all changes processed as INSERTs regardless of the DTL__CAPXACTION field,
you must code an update strategy specifying DD_INSERT in the Update Strategy Expression field.

Troubleshooting for PowerExchange Interfaces for PowerCenter 223

Index

A code pages
configuring 221
access method supported code pages 221
CAPX 18 commit processing
CAPXRT 18, 19 examples 127
Adabas in CDC sessions 161
Batch application connection attributes for sources and targets 79 target latency 127
CDC application connection attributes for sources 83 compression
Lookup relational connection attributes for sources 90 configuring 112
application name Condense
restart points 121, 179 UOW Cleanser 18, 19
configuring
code pages 221

B compression 112
encryption 112
batch extraction mode lookups for CDC data 59
PowerExchange Condense 18 lookups for IMS 58
batch mode pacing size 113
configuring sessions 133 sessions 133, 203
Before image workflows 133, 203
Flexible transactions 189 connection attributes
attributes for controlling CDC commit processing 126
configuring Maximum Rows Per commit attribute 124

C configuring PWX Latency attribute 122

configuring Real-Time Flush Latency attribute 122
cache for multiple-record writes 135 Event Table attribute 118
CAPX Image Type attribute 117
access method 18 Minimum Rows Per commit attribute 125
CAPXRT summary list of key attributes 111
access method 18, 19 connection attributes, PWXPC
CDC application connections UOW Count attribute 121
NRDB connection attributes for nonrelational sources 83 connections
DB2 connection attributes for sources 72 target connection types 64
Microsoft SQL Server connection attributes for sources 96 configuring PWXPC connections 64
Oracle connection attributes for sources 105 list by source type 62
CDC commit processing constraint-based loading
controlling with connection attributes 126 description 157
CDC data FullCBLOSupport 157
configuring lookups for 59 continuous extraction mode
group source 18, 19 PowerExchange Condense 18
CDC data map creating
extraction map 174 DB2 source definitions 29
CDC sessions DB2 target definitions 29
adding source 183 IMS source definitions 36
commit processing 161 Oracle source definitions 33
recovery example 185 VSAM source definitions 36
removing source 183 custom property
restart 121, 164, 177 FullCBLOSupport 157
restart token file 173
stopping 119, 182
Change Indicator
Flexible transactions 189
D
change mode data maps
configuring connections 121 non-relational source definitions 36
configuring sessions 146 viewing in the source definition 41

224
Datacom
Batch application connection attributes for sources and targets 79 F
CDC application connection attributes for sources 83 filelist
Lookup relational connection attributes for sources 90 description 134
datatypes Flexible Target Key
overview 217 Custom transformations 187
PowerExchange nonrelational 217
PowerExchange relational 220
transformation 217
transformation datatypes in source qualifiers 56 G
DB2 group source
CDC application connection attributes for sources 72 CDC data 18, 19
connection types 62 description 51
creating source definitions 29 multiple records 17
creating target definitions 29 sequential 17
datatypes 220 VSAM 17
DB2 Definitions 28 group target processing 53
DB2 for i5/OS
relational connection attributes for sources and targets 65
DB2 for Linux, UNIX, and Windows
relational connection attributes for sources and targets 65 H
DB2 for z/OS Hierarchy Structure metadata extension 55
relational connection attributes for sources and targets 65
TIMESTAMP datatype 220
DB2 stored procedure transformations
implementing in a mapping 60 I
DB2 Stored Procedure transformations idle time
usage considerations 59 configuring for a PWXPC session 119
DTL__CAPXACTION description 119
in CDC sessions 35, 43 IDMS
in extraction maps 47 Batch application connection attributes for sources and targets 79
DTL__CAPXRESTART1 CDC application connection attributes for sources 83
restart value 174, 175 Lookup relational connection attributes for sources 90
DTL__CAPXRESTART2 Image Type attribute
restart value 174, 175 configuring 117
DTLUAPPL importing
example 180 nonrelational source definitions 36
restart 174, 175 nonrelational target definitions 39
DTLUTSK utility IMS
description 182 CDC application connection attributes for sources 83
Lookup relational connection attributes for sources 90
Batch application connection attributes for sources and targets 79
E configuring lookups for 58
connection types 62
editing datatypes 217
extraction map definition 47 overriding the access method and related session properties 134
nonrelational source definitions 43 installing PWXPC 23
relational source definitions 35
source qualifier transformations 57
Source Qualifier transformations 57
target definitions 35, 43 L
encryption loading
configuring 112 constraints 157
enhanced restart logger token
recovery processing 184 restart value 174, 175
Event Table attribute lookup transformations 57
configuring 118 low values
extraction map preserving low values from source character fields 162
CDC data map 146
extraction map definitions 44
extraction map source definitions
editing 47 M
viewing 46 mappings overview 27
Maximum Rows Per commit attribute
configuring 124
metadata extensions
editing 43
viewing 42

Index 225
metadata extensions (continued) previewing (continued)
viewing for non-relational source definitions 47 PowerExchange nonrelational data 49
Microsoft SQL Server PowerExchange relational data 48
CDC application connection attributes for sources 96 PWX Latency attribute
relational connection attributes for sources and targets 93 configuring 122
Microsoft SQL Server definitions 31 PWXPC
Minimum Rows Per commit attribute installing 23
configuring 125
multiple-record data maps 51, 53
multiple-record writes
connection attributes 129 R
considerations for sessions 135 reader time limit
row statistics 136 description for PWXPC 72
Reader Time Limit (property)
configuring for a PWXPC session 72

N Real-Time Flush Latency attribute

configuring 122
non-relational source definitions real-time mode
editing 43 configuring sessions 146
editing metadata extensions 43 recovery
viewing data map details 41 creating the tables 173
viewing metadata extensions 47 enhanced restart 184
non-relational sources example 185
configuring batch mode sessions 138 PM_REC_STATE table 164, 165
non-relational target definitions PM_RECOVERY table 164
editing 43 PM_TGT_RUN_ID table 164
editing metadata extensions 43 state file 166
nonrelational source and target definitions 36 tables 164
nonrelational sources and targets relational connections
CDC application connection attributes for sources 83 NRDB Lookup connection attributes for nonrelational sources 90
Lookup relational connection attributes for sources 90 DB2 sources and targets 65
Batch application connection attributes for sources and targets 79 Microsoft SQL Server sources and targets 93
Oracle sources and targets 102
relational source and target definitions 28

O relational source definitions

editing 35
Oracle relational sources
CDC application connection attributes for sources 105 configuring batch mode sessions 144
connection types 62 relational target definitions
creating source definitions 33 editing 35
relational connection attributes for sources and targets 102 relational targets
Oracle definitions 33 configuring sessions 153
restart
$PMRootDir/Cache 72, 96, 105

P $PMRootDir/Restart 72, 96, 105, 121, 166

application name 121, 179
pacing size CDC sessions 121, 177
configuring 113 DTL__CAPXRESTART1 174, 175
pipeline partitioning DTL__CAPXRESTART2 174, 175
batch mode 203 DTLUAPPL 174, 175, 180
description 160, 203 DTLUAPPL example 180
loading to targets 204 earliest points 169
PowerCenter logger token 174, 175
interoperability with PowerExchange 26 null restart tokens 169
PowerExchange operation 177
performance 113 overview 164
PowerExchange Change Data Capture PM_REC_STATE table 165
Flexible transformations 189 restart token file 121, 166, 173
PowerExchange Condense restart token file folder 121
batch extraction mode 18 RESTART1 175
continuous extraction mode 18 RESTART2 175
PowerExchange Configuration File sequence token 174, 175
dbmover.cfg 25 state file 166
PowerExchange sources and targets tokens 165, 166
usage considerations 57 restart points
previewing defaults if null tokens 169
PowerExchange change data 50 earliest 169
PowerExchange data in Designer 48

226 Index
restart token file target definitions (continued)
archiving 185 editing metadata extensions 43
configuring 173 editing non-relational 43
example 176 editing relational 35
explicit override 174 viewing metadata extensions 42
special override 175 terminating conditions
syntax 173 PWPXC idle time 119
row ID PWXPC reader time limit 72
transformations affecting 157 transformations
affecting row ID 157
datatypes 56

S DB2 stored procedure 59

lookup 57
sequence fields 53, 55 update strategy 35, 43, 47
sequence token
restart value 174, 175
sequencing an dqueuing cache for multiple-record writes 135
session properties U
Library/File Override property 151 UOW Cleanser
overriding the access method and related properties for IMS bulk Condense 18, 19
sessions 134 UOW Count attribute
sessions configuring UOW Count attribute 121
considerations for multiple-record writes 135
overview 133, 203
source definitions
DB2 29 V
editing metadata extensions 43 viewing
editing, extraction maps 47 extraction map definition details 46
editing, non-relational 43 nonrelational source definition details 41
editing, relational 35 nonrelational target definition details 41
IMS 36 VSAM
viewing metadata extensions 42 Batch application connection attributes for sources and targets 79
viewing, extraction maps 46 CDC application connection attributes for sources 83
VSAM 36 connection types 62
working with non-relational source definitions 36 datatypes 217
source qualifier transformations extracting data from multiple files 134
editing 57 Lookup relational connection attributes for sources 90
source qualifiers
transformation datatypes 56
STOPTASK command
CDC sessions, stopping 119, 182 W
stored procedures workflows
transformations 59 overview 133, 203

T
target definitions
DB2 29

Index 227