Front cover

InfoSphere DataStage
for Enterprise XML
Data Integration
Addresses the complexities of
hierarchical data types

Reads huge documents using

streaming technology

Spans both batch and

real-time run times

Chuck Ballard
Vinay Bhat
Shruti Choudhary
Ravi Ravindranath
Enrique Amavizca Ruiz
Aaron Titus

ibm.com/redbooks
International Technical Support Organization

InfoSphere DataStage for Enterprise XML Data

Integration

May 2012

SG24-7987-00
 Note: Before using this information and the product it supports, read the information in
“Notices” on page ix.

First Edition (May 2012)

This edition applies to Version 8.7.1 of InfoSphere DataStage.

© Copyright International Business Machines Corporation 2012. All rights reserved.

Note to U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP
Schedule Contract with IBM Corp.
Contents

Notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
The team who wrote this book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
Now you can become a published author, too! . . . . . . . . . . . . . . . . . . . . . . . . xiv
Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv
Stay connected to IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv

Chapter 1. Heterogeneous data integration . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Importance of XML in the industry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1.1 Growth of XML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 XML business benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Information sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 XML within an organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.3 XML in new innovations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Technical benefits of XML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 DataStage as enterprise data integrator . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.1 DB2 pureXML integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.2 Real-time integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.3 Master Data Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4.4 Web Services and XML. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4.5 Real-time integration using IBM InfoSphere Change Data Capture . 18
1.4.6 Real-time integration using IBM Replication Server . . . . . . . . . . . . . 20
1.4.7 Real-time integration using Data Event Publishers . . . . . . . . . . . . . . 22
1.4.8 Real-time integration using Classic Data Event Publishers . . . . . . . 23
1.4.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Chapter 2. XML overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.1 XML introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2 The value of XML data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.3 Introduction to XML documents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.3.1 Well-formed XML documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.4 Defining document content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.4.1 Document type definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.4.2 XML schema . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.4.3 Namespaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.5 Extensible Stylesheet Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.5.1 XPath . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

© Copyright IBM Corp. 2012. All rights reserved. iii

 2.5.2 XSLT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.6 Web services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Chapter 3. Overview of IBM InfoSphere DataStage . . . . . . . . . . . . . . . . . . 45

3.1 Introduction to IBM InfoSphere Information Server . . . . . . . . . . . . . . . . . . 46
3.1.1 Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 IBM InfoSphere Information Server Architecture. . . . . . . . . . . . . . . . . . . . 50
3.3 Information Server component overview. . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3.1 Administrative clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.3.2 User clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.3.3 Server tiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.3.4 Supported topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.4 InfoSphere Information Server integration scenarios . . . . . . . . . . . . . . . . 64
3.4.1 Data integration challenges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.4.2 Building a data integration application scenario . . . . . . . . . . . . . . . . 65
3.5 Introduction to InfoSphere DataStage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.5.1 Data transformation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.5.2 Jobs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.5.3 Parallel processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.6 DataStage architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.6.1 Shared components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.6.2 Runtime architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.6.3 XML Pack architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.7 Real-time applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7.1 Always-on source stage types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7.2 End-of-wave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
3.7.3 Transaction support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.7.4 Transactional support in message-oriented applications . . . . . . . . . 86
3.8 Publishing jobs as services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.8.1 Design topology rules for always-on ISD jobs. . . . . . . . . . . . . . . . . . 93
3.8.2 Scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.8.3 Synchronizing database stages with ISD output . . . . . . . . . . . . . . . . 95
3.8.4 ISD with DTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.8.5 Information Services Director with connectors . . . . . . . . . . . . . . . . . 98
3.8.6 Partitioning in ISD jobs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.8.7 General considerations for ISD jobs . . . . . . . . . . . . . . . . . . . . . . . . 100
3.8.8 Selecting jobs for publication through ISD . . . . . . . . . . . . . . . . . . . 100

Chapter 4. XML integration in DataStage . . . . . . . . . . . . . . . . . . . . . . . . . 103

4.1 DataStage components for handling XML data. . . . . . . . . . . . . . . . . . . . 105
4.2 The XML Pack. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2.1 XML Pack Version 1.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.2.2 XML Pack Version 2.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

iv InfoSphere DataStage for Enterprise XML Data Integration

 4.2.3 XML Pack Version 3.0. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3 Web Services Pack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.4 DataStage, Information Server, and XML metadata . . . . . . . . . . . . . . . . 109
4.5 Exporting and importing DataStage repository objects . . . . . . . . . . . . . . 110

Chapter 5. Schema Library Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

5.1 Creating a library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.2 Importing an XML schema into a library . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.3 Categorizing the libraries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.4 Viewing the XML schema structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.5 Importing related schema files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.5.1 IMPORT statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.5.2 INCLUDE statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.5.3 Importing multiple schemas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.6 Errors while importing schemas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.7 Accessing the Schema Library Manager from the XML Stage . . . . . . . . 129

Chapter 6. Introduction to XML Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

6.1 The XML Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.2 Schema representation in the Assembly Editor . . . . . . . . . . . . . . . . . . . 134
6.2.1 The schema structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
6.2.2 Primitive data type representation. . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.2.3 Attribute representation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6.2.4 Representation of repeating elements (LIST) . . . . . . . . . . . . . . . . . 136
6.2.5 DataStage links representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
6.2.6 Complex-type representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.2.7 Representation of recursive elements. . . . . . . . . . . . . . . . . . . . . . . 140
6.2.8 Representation of derived types . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.2.9 Representation of an element with all optional elements . . . . . . . . 142
6.3 The assembly model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

Chapter 7. Consuming XML documents . . . . . . . . . . . . . . . . . . . . . . . . . . 147

7.1 XML Parser step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.1.1 Reading XML data from a database . . . . . . . . . . . . . . . . . . . . . . . . 149
7.1.2 Reading multiple XML files from a single directory or folder . . . . . . 152
7.1.3 Reading a single file directly from the XML Stage . . . . . . . . . . . . . 154
7.1.4 Specifying the XML schema in the Document Root tab . . . . . . . . . 156
7.1.5 Validating the XML file against the schema . . . . . . . . . . . . . . . . . . 159
7.1.6 The Reject action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
7.1.7 Treating all XML data types as string . . . . . . . . . . . . . . . . . . . . . . . 167
7.1.8 The output of the XML Parser step . . . . . . . . . . . . . . . . . . . . . . . . . 168
7.2 The Output step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
7.2.1 Automatically mapping the columns . . . . . . . . . . . . . . . . . . . . . . . . 176
7.2.2 Propagating columns from the source to the target . . . . . . . . . . . . 177

Contents v
 7.2.3 Mapping constants to a target column . . . . . . . . . . . . . . . . . . . . . . 178
7.2.4 Selecting mapping candidates by using the More option . . . . . . . . 181
7.3 Mapping suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
7.3.1 Mapping child and parent items to same output list . . . . . . . . . . . . 183
7.3.2 Mapping when no lists are available in the Parser step . . . . . . . . . 185
7.4 Parsing large schemas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
7.4.1 Output schema location of the second XML Parser step . . . . . . . . 190
7.4.2 Configuring the auto-chunked elements in the assembly . . . . . . . . 193
7.4.3 Designing the assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
7.5 Parsing only a section of the schema . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
7.6 Incorporating an XSLT style sheet during parsing . . . . . . . . . . . . . . . . . 198

Chapter 8. Creating and composing XML documents . . . . . . . . . . . . . . . 201

8.1 The XML Composer step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.1.1 Writing XML files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.1.2 Specifying the XML schema in the Document Root tab . . . . . . . . . 209
8.1.3 Creating a valid XML file . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
8.1.4 The Mappings tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
8.1.5 The Header tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
8.1.6 The Format tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
8.1.7 The output of the Composer step . . . . . . . . . . . . . . . . . . . . . . . . . . 230
8.2 Creating the hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
8.2.1 The Regroup step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.2.2 The HJoin step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
8.2.3 Regroup step versus HJoin step . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
8.2.4 H-Pivot step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
8.3 Multiple XML documents: Single Composer step . . . . . . . . . . . . . . . . . . 253
8.3.1 Target tab configuration for creating multiple XML documents . . . . 255
8.4 Configuring the Output step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
8.5 Creating XML files for various kinds of schemas . . . . . . . . . . . . . . . . . . 259
8.5.1 Creating an XML file for a schema with a choice definition. . . . . . . 259
8.5.2 Creating the XML file for a schema with type derivatives . . . . . . . . 263

Chapter 9. Transforming XML documents . . . . . . . . . . . . . . . . . . . . . . . . 265

9.1 The Sort step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
9.2 The Aggregate step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
9.3 The Switch step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
9.4 The Union step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
9.5 The V-Pivot step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
9.6 The OrderJoin step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

Chapter 10. DataStage performance and XML . . . . . . . . . . . . . . . . . . . . . 291

10.1 Partitioning and collecting methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
10.1.1 Keyless partitioning methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294

vi InfoSphere DataStage for Enterprise XML Data Integration

 10.1.2 Keyed partitioning methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
10.1.3 Collectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
10.1.4 Choosing a collector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
10.2 Sorting relational data basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
10.3 Sorting relational data for composing XML documents. . . . . . . . . . . . . 299
10.4 Partitioning and parallel processing with XML documents . . . . . . . . . . 304
10.4.1 Using DataStage partition parallelism . . . . . . . . . . . . . . . . . . . . . . 305
10.4.2 Parallel parsing large XML documents . . . . . . . . . . . . . . . . . . . . . 310
10.4.3 Advanced parallel parsing strategy . . . . . . . . . . . . . . . . . . . . . . . . 316
10.5 Tuning XML Transformation steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
10.5.1 Sort step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
10.5.2 H-Join step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
10.6 Passing XML documents or chunks within the job . . . . . . . . . . . . . . . . 320
10.6.1 Passing XML in-line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
10.6.2 Passing XML by reference as a large object column . . . . . . . . . . 321
10.6.3 Managing large XML schemas with LOB references . . . . . . . . . . 324
10.7 Storing XML intermediately on disk. . . . . . . . . . . . . . . . . . . . . . . . . . . . 327

Chapter 11. Developing XML solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . 329

11.1 Integration with Web services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
11.1.1 Importing WSDL and XML schema information . . . . . . . . . . . . . . 330
11.1.2 Designing a DataStage job . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
11.2 Information Services Director: DataStage jobs as services . . . . . . . . . 340
11.2.1 Understanding how end-of-wave works . . . . . . . . . . . . . . . . . . . . 340
11.2.2 Designing and testing DataStage jobs with end-of-wave . . . . . . . 341
11.2.3 Expected Peek stage output when parsing a document . . . . . . . . 341
11.2.4 Expected Peek stage output when composing a document . . . . . 342
11.3 Event-driven processing with IBM WebSphere MQ . . . . . . . . . . . . . . . 342
11.4 XML data integration with relational databases. . . . . . . . . . . . . . . . . . . 344
11.4.1 IBM DB2 V9 and DB2 pureXML . . . . . . . . . . . . . . . . . . . . . . . . . . 344
11.4.2 Oracle 11g and Oracle XML DB . . . . . . . . . . . . . . . . . . . . . . . . . . 345
11.4.3 Performance considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
11.5 Migrating from XML Pack 2.0 for DataStage . . . . . . . . . . . . . . . . . . . . . 349
11.5.1 XML document validation and reject functionality . . . . . . . . . . . . . 349
11.5.2 Design patterns for composing XML documents . . . . . . . . . . . . . 354

Appendix A. Supported type mappings . . . . . . . . . . . . . . . . . . . . . . . . . . 363

A.1 DataStage data type to XML data type mapping . . . . . . . . . . . . . . . . . . 364
A.2 XML data type to DataStage data type mapping . . . . . . . . . . . . . . . . . . 365
A.3 Connector stage LOB reference support matrix . . . . . . . . . . . . . . . . . . . 367

Appendix B. Transformer functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371

Contents vii
 Abbreviations and acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Online resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Help from IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

viii InfoSphere DataStage for Enterprise XML Data Integration

Notices

This information was developed for products and services offered in the U.S.A.

IBM may not offer the products, services, or features discussed in this document in other countries. Consult
your local IBM representative for information on the products and services currently available in your area.
Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM
product, program, or service may be used. Any functionally equivalent product, program, or service that
does not infringe any IBM intellectual property right may be used instead. However, it is the user's
responsibility to evaluate and verify the operation of any non-IBM product, program, or service.

IBM may have patents or pending patent applications covering subject matter described in this document.
The furnishing of this document does not give you any license to these patents. You can send license
inquiries, in writing, to:
IBM Director of Licensing, IBM Corporation, North Castle Drive, Armonk, NY 10504-1785 U.S.A.

The following paragraph does not apply to the United Kingdom or any other country where such
provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION
PROVIDES THIS PUBLICATION "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR
IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT,
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer
of express or implied warranties in certain transactions, therefore, this statement may not apply to you.

This information could include technical inaccuracies or typographical errors. Changes are periodically made
to the information herein; these changes will be incorporated in new editions of the publication. IBM may
make improvements and/or changes in the product(s) and/or the program(s) described in this publication at
any time without notice.

Any references in this information to non-IBM websites are provided for convenience only and do not in any
manner serve as an endorsement of those websites. The materials at those websites are not part of the
materials for this IBM product and use of those websites is at your own risk.

IBM may use or distribute any of the information you supply in any way it believes appropriate without
incurring any obligation to you.

Information concerning non-IBM products was obtained from the suppliers of those products, their published
announcements or other publicly available sources. IBM has not tested those products and cannot confirm
the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on
the capabilities of non-IBM products should be addressed to the suppliers of those products.

This information contains examples of data and reports used in daily business operations. To illustrate them
as completely as possible, the examples include the names of individuals, companies, brands, and products.
All of these names are fictitious and any similarity to the names and addresses used by an actual business
enterprise is entirely coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, which illustrate programming
techniques on various operating platforms. You may copy, modify, and distribute these sample programs in
any form without payment to IBM, for the purposes of developing, using, marketing or distributing application
programs conforming to the application programming interface for the operating platform for which the
sample programs are written. These examples have not been thoroughly tested under all conditions. IBM,
therefore, cannot guarantee or imply reliability, serviceability, or function of these programs.

© Copyright IBM Corp. 2012. All rights reserved. ix

Trademarks
IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business
Machines Corporation in the United States, other countries, or both. These and other IBM trademarked
terms are marked on their first occurrence in this information with the appropriate symbol (® or ™),
indicating US registered or common law trademarks owned by IBM at the time this information was
published. Such trademarks may also be registered or common law trademarks in other countries. A current
list of IBM trademarks is available on the Web at http://www.ibm.com/legal/copytrade.shtml

The following terms are trademarks of the International Business Machines Corporation in the United States,
other countries, or both:

AIX® InfoSphere® Redbooks (logo) ®

alphaWorks® LoadLeveler® solidDB®
DataStage® MQSeries® System z®
DB2® Orchestrate® Tivoli®
developerWorks® pureXML® WebSphere®
IBM® QualityStage® z/OS®
IMS™ Redbooks®

The following terms are trademarks of other companies:

Netezza, and N logo are trademarks or registered trademarks of IBM International Group B.V., an IBM
Company.

Linux is a trademark of Linus Torvalds in the United States, other countries, or both.

Microsoft, Windows, and the Windows logo are trademarks of Microsoft Corporation in the United States,
other countries, or both.

Java, and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its
affiliates.

UNIX is a registered trademark of The Open Group in the United States and other countries.

Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo, Celeron, Intel Xeon, Intel
SpeedStep, Itanium, and Pentium are trademarks or registered trademarks of Intel Corporation or its
subsidiaries in the United States and other countries.

Other company, product, or service names may be trademarks or service marks of others.

x InfoSphere DataStage for Enterprise XML Data Integration

Preface

XML is one of the most common standards for the exchange of information.
However, organizations find challenges in how to address the complexities of
dealing with hierarchical data types, particularly as they scale to gigabytes and
beyond. In this IBM® Redbooks® publication, we discuss and describe the new
capabilities in IBM InfoSphere® DataStage® 8.5. These capabilities enable
developers to more easily manage the design and processing requirements
presented by the most challenging XML sources. Developers can use these
capabilities to create powerful hierarchical transformations and to parse and
compose XML data with high performance and scalability. Spanning both batch
and real-time run times, these capabilities can be used to solve a broad range of
business requirements.

As part of the IBM InfoSphere Information Server 8.5 release, InfoSphere

DataStage was enhanced with new hierarchical transformation capabilities called
XML Stage. XML Stage provides native XML schema support and powerful XML
transformation functionality. These capabilities are based on a unique
state-of-the-art technology that allows you to parse and compose any complex
XML structure from and to a relational form, as well as to a separate hierarchical
form.

Also, XML Stage has the capability to read single huge documents using a new
streaming methodology that avoids the need to load the document into memory.
XML Stage has support for any type of XSD, or a collection of XSDs, to define
your XML metadata. The most important capability is a whole new hierarchical
editing mode called an assembly, which provides support for the creation of
complex multi-node hierarchical structures. Much more function exists, such as
the explicit control of XML validation, a built-in test facility to ease transformation
development, and support for both enterprise edition (EE) and server jobs.

This book is targeted at an audience of systems designers and developers who

focus on implementing XML integration support in their environments. It includes
information about the following subjects:
򐂰 Overview of IBM DataStage and XML technology, features, and benefits
򐂰 Integration of DataStage and XML and how they support the various
platforms and environments
򐂰 Parsing, transforming, and composing XML documents
򐂰 Parallel processing and performance tuning of the environment
򐂰 Developing XML solutions, including migration and integration scenarios

© Copyright IBM Corp. 2012. All rights reserved. xi

The team who wrote this book
This book was produced by a team of specialists from around the world working
with the International Technical Support Organization, in San Jose, California.
We list the team members, along with a short biographical sketch of each team
member.

Chuck Ballard is a Project Manager at the International

Technical Support Organization, in San Jose, California. He
has over 35 years experience, holding positions in the areas
of Product Engineering, Sales, Marketing, Technical
Support, and Management. His expertise is in the areas of
database, data management, data warehousing, business
intelligence, and process re-engineering. He has written
extensively on these subjects, taught classes, and presented
at conferences and seminars worldwide. Chuck has both a
Bachelors degree and a Masters degree in Industrial Engineering from Purdue
University.

Vinay Bhat is a Technical Sales Specialist, and a member of

the sales team for IBM Information Management software
products. He has over eight years of experience in
architecture, design, implementation, and management of
Data Integration solutions using the IBM Information Server
and information management products. His experience
includes extensive development of IBM Information Server
DataStage, enterprise data warehouse solutions, information
governance, and MDM solutions. Vinay has helped develop
proven practices and deliver workshops to facilitate successful client
deployments using IBM Information Server.

Shruti Choudhary is a Senior Quality Assurance (QA)

Engineer at the IBM Software labs in India, with over three
years of experience. She has extensive experience with
Information Server having deep-rooted expertise on
DataStage XML Stage. Shruti has written several IBM
developerWorks® articles and has given several demos and
presentations to clients on XML Stage. She has a Bachelors
degree in Electronic and Communication Engineering from
Visvesvaraya Technological University.

xii InfoSphere DataStage for Enterprise XML Data Integration

 Ravi Ravindranath is a Senior Solutions Architect at the
IBM Information Management Technology Ecosystem in
Dallas, Texas, specializing in InfoSphere information
integration middleware products. He has over 30 years of
experience in lead positions in the areas of software
development, product management, and pre-sales. Ravi has
implemented several turn-key projects for major airlines,
telecommunications, financial institutions, federal and state
governments, and has contributed to the five-year IT strategy
plans of the Government of Indonesia and the Government of India. He advises
IBM Business Partners, mentors field sales, and presents at conferences and
seminars worldwide. Ravi has a Bachelors degree in Engineering from the
University of Madras and an MBA from the University of Dallas in Texas.

Enrique Amavizca Ruiz is an Information Server

Consultant at the Information Management Technology
Ecosystem (IMTE) in Mexico D.F. He has over 16 years of
experience in positions in the areas of development,
technical support, systems administration, management,
and consulting. His expertise is in the areas of database,
data integration, and software development. He has worked
extensively on these subjects, and taught Information Server
courses in North and South America. Enrique holds a
Bachelors degree in Computer Systems Engineering from Universidad de las
Americas-Puebla.

Aaron Titus is an Advanced Technical Support Engineer for

the IBM InfoSphere Platform products, located in Littleton,
Massachusetts. He has provided technical support for
Information Server, DataStage, and other data integration
products since 1999. Aaron has taught classes to clients and
IBM Business Partners, and presented workshops and
demos at user conferences. His expertise is with data
integration technology, specializing in parallel processing,
clustering and GRID deployments, real time, and
service-oriented architectures (SOA).

Other contributors
In this section, we thank others who contributed to this book, in the form of
written content, subject expertise, and support.

Preface xiii
 From IBM locations worldwide
Ernie Ostic - Client Technical Specialist, IBM Software Group, Worldwide Sales
Enablement, Piscataway, NJ
Tony Curcio - InfoSphere Product Management, Software Product Manager,
Charlotte, NC

From the International Technical Support Organization

Mary Comianos - Publications Management
Ann Lund - Residency Administration
Emma Jacobs - Graphics Support

Now you can become a published author, too!

Here’s an opportunity to spotlight your skills, grow your career, and become a
published author—all at the same time! Join an ITSO residency project and help
write a book in your area of expertise, while honing your experience using
leading-edge technologies. Your efforts will help to increase product acceptance
and customer satisfaction, as you expand your network of technical contacts and
relationships. Residencies run from two to six weeks in length, and you can
participate either in person or as a remote resident working from your home
base.

Find out more about the residency program, browse the residency index, and
apply online at:
ibm.com/redbooks/residencies.html

Comments welcome
Your comments are important to us!

We want our books to be as helpful as possible. Send us your comments about

this book or other IBM Redbooks publications in one of the following ways:
򐂰 Use the online Contact us review Redbooks form found at:
ibm.com/redbooks
򐂰 Send your comments in an email to:
[email protected]
򐂰 Mail your comments to:

xiv InfoSphere DataStage for Enterprise XML Data Integration

 IBM Corporation, International Technical Support Organization
Dept. HYTD Mail Station P099
2455 South Road
Poughkeepsie, NY 12601-5400

Stay connected to IBM Redbooks

򐂰 Find us on Facebook:
http://www.facebook.com/IBMRedbooks
򐂰 Follow us on Twitter:
http://twitter.com/ibmredbooks
򐂰 Look for us on LinkedIn:
http://www.linkedin.com/groups?home=&gid=2130806
򐂰 Explore new Redbooks publications, residencies, and workshops with the
IBM Redbooks weekly newsletter:
https://www.redbooks.ibm.com/Redbooks.nsf/subscribe?OpenForm
򐂰 Stay current on recent Redbooks publications with RSS Feeds:
http://www.redbooks.ibm.com/rss.html

Preface xv
xvi InfoSphere DataStage for Enterprise XML Data Integration
 1

Chapter 1. Heterogeneous data

integration
Most key business initiatives require the integration of data from heterogeneous
data sources. These initiatives can be accomplished by developing a custom
solution for each data source, as depicted in Figure 1-1 on page 2. In such an
architecture, as the number of data sources increases, the complexity required to
integrate them increases. Costs for managing the infrastructure skyrocket. Using
the eXtensible Markup Language (XML) to standardize the exchange of
information between disparate sources is becoming a modern, industry standard
solution. This book addresses the XML technology and the use of IBM
InfoSphere DataStage to successfully delivery enterprise-wide heterogeneous
data integration.

In this book, we discuss in detail the creation of Information Server DataStage

jobs that integrate disparate data using XML technology. You can understand the
topics in this book better if you have a basic working knowledge of InfoSphere
DataStage, and general understanding of XML concepts. You can obtain more
information about how to develop DataStage jobs, and reference information
about XML technologies in “Related publications” on page 375.

© Copyright IBM Corp. 2012. All rights reserved. 1

Figure 1-1 Heterogeneous data integration

Enterprise data integration initiatives present many challenges. You might deploy
enterprise services using a service-oriented architecture (SOA), implementing a
cloud initiative, or building the infrastructure for Dynamic Data Warehousing. For
many of these tasks, XML can be an important building block for your integration
architecture.

Implementing innovative solutions using XML technology is an integral part of

every enterprise. The XML Transformation capabilities of InfoSphere DataStage
can handle large, complex XML documents and process thousands of real-time
transactions with Web Services or message bus implementations. XML-based
solutions offer high flexibility and ease of adaptation, making them useful in a
wide spectrum of industries and business solutions. In this book, we highlight
several of the most common XML-based solutions, and how to use IBM
InfoSphere DataStage to successfully implement these solutions in your
enterprise.

2 InfoSphere DataStage for Enterprise XML Data Integration

1.1 Importance of XML in the industry
XML technology is pervasive in virtually all industries and sectors, because of its
versatility and neutrality for exchanging data among diverse devices,
applications, and systems from various vendors, as depicted in Figure 1-2.
These qualities of XML and its easy to understand self-describing nature, ability
to handle structured, semi-structured, and unstructured data, and support for
Unicode make XML a universal standard for data interchange.

App1 CSV

ASCII App2
XML
WH
DB

Appn

Figure 1-2 Data exchange with XML

1.1.1 Growth of XML

Nearly every company today comes across XML in some form. The amount of
XML data with which organizations work is growing at a rapid rate. In fact, it is
estimated that the volume of XML data is growing twice as fast as that of the
traditional data that typically resides in relational databases. The following factors
fuel the growth of XML data:
򐂰 XML-based industry and data standards
򐂰 Service-oriented architectures (SOAs) and Web Services
򐂰 Web 2.0 technologies, such as XML feeds and syndication services

XML-based industry standards

Almost every industry has multiple standards based on XML, and there are
numerous cross-industry XML standards, as well.

Chapter 1. Heterogeneous data integration 3

 The following list includes a few examples of XML-based industry standards:
򐂰 ACORD - XML for the Insurance Industry:
http://www.acord.org/
򐂰 FPML - Financial Product:
http://www.fpml.org/
򐂰 HL7 - Health Care:
http://www.hl7.org/
򐂰 IFX - Interactive Financial Exchange:
http://www.ifxforum.org/
򐂰 IXRetail - Standard for Retail operation:
http://www.nrf-arts.org/
򐂰 XBRL - Business Reporting/Accounting:
http://www.xbrl.org/
򐂰 NewsML - News/Publication:
http://www.newsml.org/

These standards facilitate purposes, such as the exchange of information

between the various players within these industries and their value chain
members, data definitions for ongoing operations, and document specifications.
More and more companies adopt XML standards or are compelled to adopt them
to stay competitive, improve efficiencies, communicate with their trading partners
or suppliers, or perform everyday tasks.

1.2 XML business benefits

The benefits of a common base upon which the technical world can build layers
upon layers of technical innovation are enormous. These benefits are only
possible if this standard is agreed to by all.

The following list shows the primary XML applications:

򐂰 Sharing information: The main problem with integrating data between any two
business organizations is the interface between them. If they can at least
agree upon the standard of their common meeting point and its usage, they
can build upon this standard to start building their applications. If there is
already an existing interface or infrastructure provided by industry or
government standard or infrastructure, the business cost of developing it is
extinguished.

4 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 Storage, transmission, and interpretation of data: If the storage of information
is adaptable to many mediums, its cost is driven to the lowest cost of all
mediums. XML is based on text, which obviously is accepted by all. It can be
read. Its storage is cheap, compared to the storage requirements of graphics.
And, because the size of this text-based object is small, its transmission, not
withstanding cost, is inexpensive as well. And, it is commonly accepted,
because it adheres to worldwide standards and it is easily interpreted.
򐂰 Security: With the explosion of confidential information transmitted across the
Internet, sophisticated levels of security are now necessary. Companies need
to protect their documents and email, banks need to allow their depositors to
download their accounts, and merchants need to be available for their
customers to enter their credit card details without compromising their
security and privacy. The more secure the transmission medium, the more
confidentiality it provides, and the more it can be used to advertise a
competitive advantage. With the evolution of XML digital signatures, the ability
to protect or hide parts of a document, while sitting on a PC, server or
mainframe, now covers up a security patch.
򐂰 Speed and the amount of content delivery: With the rapid evolution of network
technologies, the speed and delivery of any content continue to grow in
importance as a product evaluation criteria. Network companies now
advertise download times of movies and CDs. Again, the first companies
discovering the abilities of delivering something at an increasing speed
without compromising their content gains the competitive advantage.

1.2.1 Information sharing

XML has been readily accepted by the technical world because of its simplicity.
The benefits of having a common format to share information between any two
organizations are obvious. Technologies and standards have been built upon
XML. Consortiums and business organizations have developed industry-wide
XML formats and standardization. Examples are inter-bank payments between
banks, insurance companies, and trading agencies; supply chains between
manufacturers, distributors, and retailers; and battlefield information between
soldiers, satellites, and defense analysts.

In December 2000, the United Nations Centre for Trade Facilitation and
Electronic Business (UN/CEFACT) and the Organizations for the Advancements
of Standard Information Standards (OASIS) came together to initiate a project to
standardize XML specifications for business. This initiative, which is called the
Electronic Business XML (ebXML), developed a technical framework that
enabled XML to be utilized for all exchange of all electronic business data. The
main aim of ebXML was to lower the cost and difficulties, focusing on small and
medium-sized business and developing nations, in order to facilitate international
trade.

Chapter 1. Heterogeneous data integration 5

 Another example, BizTalk, is an initiative supported by Microsoft, SAP, and
Boeing, among other major vendors. However, BizTalk is not a standard. BizTalk
is more of a community of users. It aims to enable the consistent adoption of
XML to ease use and, therefore, easily adopt electronic commerce and
application integration.

1.2.2 XML within an organization

With the emergence of Customer Relationship Management (CRM) and
Enterprise Architecture Integration (EAI), client-oriented organizations
re-engineered their systems to provide a whole new experience for the
consumer.

CRM and EAI involve bringing together a multitude of applications and systems
from multiple vendors to behave as a coordinated whole. The common thread
among all these applications is XML. Imagine trying to integrate information from
billing histories, customer personal information and credit histories, distribution
systems, and workforce management systems; and then displaying them using
browsers, databases, processes, and workflow systems on a variety of technical
platforms, such as mainframes, PCs, and medium-sized servers. All this
information can be shared by using XML, because it is a versatile, powerful, and
flexible language. It has the ability to describe complex data. It is also extensible,
allowing applications to grow and develop without architectural re-engineering.

IBM uses XML in all its new tools. The IBM applications WebSphere® Studio,
DB2®, and WebSphere Application Server are based on XML with extensibility
as a major advantage.

1.2.3 XML in new innovations

The innovations based on XML are too numerous to list. So many new ideas are
now based in XML that entrepreneurial companies will find it difficult to avoid the
use of XML and its standards.

Voice XML
In October, 2001, the World Wide Web Consortium (W3C) announced the first
release of a working draft for the Voice Extensible Markup Language
(VoiceXML). VoiceXML has these purposes among others:
򐂰 Hides designers and developers from low-level platform-specific details.
򐂰 Promotes portability across implementation platforms.
򐂰 Offers a common language for platform providers, development tool
providers, and content providers.

6 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 Provides features to support complex dialogs, yet it is easy to use for simple
interactions.
򐂰 Caters to audio dialogs that feature digitized audio, speech recognition,
telephony, and synthesized speech. Its goal is to deliver the advantages of
web-based development and context delivery for interactive voice response
applications

Scalable Vector Graphics

Scalable Vector Graphics (SVG) is an XML-based language for describing
two-dimensional graphics. Its main use is in Geographical Information Systems
(GIS) where travel maps, council boundaries, forest fires, and natural disasters
can be processed and displayed independently of technical platforms. It is able to
integrate with non-GIS software, and also allow graphic elements to be
non-graphic.

The benefits of this new XML derivative are not difficult to comprehend. Its uses
are many. For example, insurance companies can estimate and forecast natural
disaster claims. Scientists can study environmental impacts. And, local and
federal governments can perform city and town planning.

1.3 Technical benefits of XML

The basics of XML do not change. What changed is the extent that XML is used
today and how it is incorporated into technologies for the future. XML offers many
benefits, several of which are listed.

Acceptability of use for data transfer

XML is not a programming language. It is a standard way of putting information in
a format that can be processed and exchanged across hardware devices,
operating systems, software applications, and the web. It is such a common
medium of data that it enables the transmission and retrieval, and storage of
information over the Internet across company boundaries, making it a natural
choice for data management for e-business transactions.

Uniformity and conformity

The inability of two computer systems or applications to talk to each other is
always a challenge. When two applications are integrated, the business and
technical experts must decide either to integrate the two systems, or to
re-architect the applications. If data from both applications conforms to a format
and is easily transformed from one to another, development costs can be
reduced. If this common format can be further developed and is accepted
industry-wide, interfacing the applications to other applications is less costly.

Chapter 1. Heterogeneous data integration 7

 Simplicity and openness
Information that is coded in XML is visually read and accepted. Because it can be
processed by computers easily, XML is widely accepted by major vendors in the
computing world. Microsoft has indicated that it will use XML as the exchange
format for its Microsoft Office software suite. Both Microsoft and Netscape web
browsers support XML.

XML garners interest, because it is simple to understand, implement, and use. It

follows the Pareto principle, an 80/20 solution, meaning that it supplies about
80% of the functionality of competing technologies with perhaps 20% of the effort
that is required to build enterprise-level solutions.

XML is not a total solution for every problem in e-business, but it is making
significant inroads in communications between old computer programs.
Therefore, these old programs last longer, saving money and time, which are
important when both are so precious to the bottom line.

Separation of data and display

Separation of data and its display is not a new paradigm in the computing world.
In the last few years, application designers have raised the concept of the
Model-View-Controller (MVC) approach to building applications for many
reasons. Without separation of data, re-use of that data in multiple user
interfaces is difficult. Websites evolved radically over the last eight years. Every
year, websites must be upgraded to compete for consumer attention. Websites
need better attention-getting displays and response times. If the data must be
rehashed every time that an upgrade to the website is required, the development
costs increase. However, if you can reuse or build upon existing data, you can
save on redevelopment costs.

Imagine a navigation system that is used by consumers to move from one place
to another. This system has street maps, address information, local attractions,
and other information. If this information must be displayed on a web browser,
personal device assistant (PDA), or a mobile phone, it is a major development
cost if we must develop three data access systems and three data presentation
systems. However, if we develop one data access system, we also need to
create three data presentation files for each system. We transform XML using the
Extensible Stylesheet Language Transformation (XSLT) feature.

Extensibility
HTML has a major problem in that it is not extensible. It is enhanced by software
vendors. These enhancements are not coordinated and, therefore, non-standard.
It was never designed to access data from databases. To overcome this
deficiency, Microsoft built Active Server Pages (ASP) and Sun produced
JavaServer Pages (JSP).

8 InfoSphere DataStage for Enterprise XML Data Integration

 XML, however, was designed initially to allow extensions.

Industry acceptance
XML is accepted widely by the information and computing industry. It is based on
common concepts. A large number of XML tools will emerge from both
existing software vendors and XML start-up companies. XML is readable by
every operating system, because it is in ASCII text. XML can be seen by any
text editor or word processor.

The tree-based structure of XML is much more powerful than fixed-length data
formats. Because objects are tree structures as well, XML is ideally suited to
working with object-oriented programming.

1.4 DataStage as enterprise data integrator

IBM InfoSphere Information Server is a software platform that helps
organizations derive more value from the complex, heterogeneous information
that is spread across their systems. It provides breakthrough collaboration,
productivity, and performance for cleansing, transforming, and moving this
information consistently and securely throughout the enterprise, as depicted in
Figure 1-3 on page 10. It can then be accessed and used in new ways to drive
innovation, increase operational efficiency, and lower risk.

As a component of IBM InfoSphere Information Server, IBM InfoSphere

DataStage integrates data across multiple, high-volume data sources and target
applications. It integrates data on demand with a high-performance parallel
framework, extended metadata management, and enterprise connectivity.
DataStage supports the collection, integration, and transformation of large
volumes of data, with data structures ranging from simple to highly complex.

Chapter 1. Heterogeneous data integration 9

 Figure 1-3 DataStage as data integrator

DataStage can manage data arriving in real time, as well as data received on a
periodic or scheduled basis, which enables companies to solve large-scale
business problems through high-performance processing of massive data
volumes. By making use of the parallel processing capabilities of multiprocessor
hardware platforms, IBM InfoSphere DataStage Enterprise Edition can scale to
satisfy the demands of ever-growing data volumes, stringent real-time
requirements, and ever-shrinking batch windows.

Along with these key components, establishing consistent development

standards helps to improve developer productivity and reduce ongoing
maintenance costs. Development standards can also make it easier to integrate

10 InfoSphere DataStage for Enterprise XML Data Integration

 external processes, such as automated auditing and reporting, and to build
technical and support documentation.

1.4.1 DB2 pureXML integration

With the introduction of DB2 9 to the market, IBM pureXML® is a new feature
that provides the capability of storing XML data natively in a database table.
Before DB2 9, management of XML data involved these common approaches:
򐂰 Storing XML documents on file systems
򐂰 Stuffing XML data into large objects (LOBs) in relational databases
򐂰 Shredding XML data into multiple relational columns and tables
򐂰 Isolating data into XML-only database systems

Frequently, these obvious choices for managing and sharing XML data do not
meet performance requirements. File systems are fine for simple tasks, but they
do not scale well when you have hundreds or thousands of documents.
Concurrency, recovery, security, and usability issues become unmanageable.

With the release of DB2 9, IBM is leading the way to a new era in data
management. DB2 9 embodies technology that provides pure XML services.
This pureXML technology is not only for data server external interfaces; rather,
pureXML extends to the core of the DB2 engine. The XML and relational
services in DB2 9 are tightly integrated. They offer the industry’s first pureXML
and relational hybrid data server. Figure 1-4 illustrates the hybrid database.

Integrated Information
XQuery SQL
Language
Flexibility
Common
Services

<XML>
Optimized
Storage

RDB

</XML>

Figure 1-4 pureXML and relational hybrid database

Chapter 1. Heterogeneous data integration 11

 The pureXML technology in DB2 9 includes the following capabilities:
򐂰 pureXML data type and storage techniques for the efficient management of
hierarchical structures common in XML documents
򐂰 pureXML indexing technology to speed up searches of subsets of XML
documents
򐂰 New query language support (for XQuery and SQL/XML) based on industry
standards and new query optimization techniques
򐂰 Industry-leading support for managing, validating, and evolving XML schemes
򐂰 Comprehensive administrative capabilities, including extensions to popular
database utilities
򐂰 Integration with popular application programming interfaces (APIs) and
development environments
򐂰 XML shredding and publishing facilities for working with existing relational
models
򐂰 Enterprise-proven reliability, availability, scalability, performance, security, and
maturity that you expect from DB2

1.4.2 Real-time integration

Much of the DataStage history focuses on batch-oriented applications. Many
mechanisms and techniques were devised to process huge amounts of data in
the shortest amount of time, in the most efficient way, in that environment. We
are aware that concerns exist relating to limiting the number of job executions,
processing as much data as possible in a single job run, and optimizing the
interaction with the source and target databases (by using bulk SQL techniques).
All these concerns relate to bulk data processing. Batch applications tend to be
bound to time constraints, such as batch execution windows, which typically
require at least a few hours for daily processing. However, several applications do
not process huge amounts of data at one time, but rather process several small,
individual requests. The number and size of requests are not minimal. They
might actually be relatively large, but not as large as high-volume batch
applications. For high-volume scenarios, users typically resort to batch-oriented
techniques. The types of applications in this section focus on scenarios with a
large number of requests, each of varying size, spanning a long period of time.

The following list describes these types of applications:

򐂰 Real-time transactions:
– Business-to-consumer applications: A person waits for a response, such
as in a web-based or online application.
– Business-to-business applications: The caller is another application.

12 InfoSphere DataStage for Enterprise XML Data Integration

򐂰 Real-time reporting/decision making:
– Few or no updates to database tables occur.
– Fetching and massaging large amounts of data are involved.
– The application generates support for a specific decision, such as an
insurance approval.
򐂰 Real-time data quality:
– Other applications invoke IBM InfoSphere QualityStage® for cleansing
and de-duplication, for example, the use of Master Data Management
(MDM). See 1.4.3, “Master Data Management” on page 15.
򐂰 Integration with third-party business process control systems
A business process coordinator orchestrates the processing of a series of
steps, involving multiple participants. One of the participants might be
DataStage, which receives requests and posts results back to the
orchestrator through Java Message Services (JMS), for example. A
third-party example is TIBCO.
򐂰 Near-real time:
– Applications are message-oriented for instant access and update.
– Data trickles from a source mainframe to populate active data warehouses
and reporting systems.

Two capabilities were added to DataStage over time to address the needs of real
time:
򐂰 Message Queue/Distributed Transaction Stage (MQ/DTS): MQ/DTS
addresses the need for guaranteed delivery of source messages to target
databases, with the once-and-only-once semantics. This type of delivery
mechanism was originally made available in DataStage 7.5 in the form of the
Unit-of-Work (UOW) stage. The original target in DS 7.5 was Oracle. In
InfoSphere DataStage 8.x, this solution is substantially upgraded
(incorporating the new database connector technology for various types of
databases) and rebranded as the Distributed Transaction Stage.
򐂰 Information Services Director (ISD): ISD enables DataStage to expose DS
jobs as services for service-oriented applications (SOA). ISD supports, as
examples, the following types of bindings:
– SOAP
– Enterprise JavaBeans (EJB)
– JMS

Chapter 1. Heterogeneous data integration 13

 DTS and ISD work differently and serve separate purposes. In this section, we
describe the specifics of each capability. However, common aspects relate to
both capabilities:
򐂰 Job topologies: Real-time jobs of any sort need to obey certain rules so they
can operate in a request/response fashion.
򐂰 Transactional support:
– Most real-time applications require updating a target database. Batch
applications can tolerate failures by restarting jobs, as long as the results
at the end of processing windows are consistent.
– Real-time applications cannot afford the restarting of jobs. For each
request, the net result in the target database must be consistent.
򐂰 End-of-wave:
– The parallel framework implements virtual datasets (memory buffers) that
are excellent for batch applications. Virtual datasets are a key optimization
mechanism for high-volume processing.
– Real-time applications cannot afford records sitting in buffers waiting to be
flushed.
– The framework was adapted to support end-of-wave, which forces the
flushing of memory buffers so responses are generated or transactions
committed.
򐂰 Payload processing:
– Frequently, batch applications deal with large payloads, such as big XML
documents.
– Common payload formats are COBOL and XML.
򐂰 Pipeline parallelism challenges: Fundamental for performance, the concept of
pipeline parallelism introduces challenges in real time, but those challenges
can be circumvented.

Definition of real time

The concept of real time is not consistent with applications that are typically used
for embedded systems, such as avionics, traffic control, and industrial control.
Similarly, DataStage is typically not used for embedded applications or for
scenarios in which there are stringent timing constraints.

The term real time applies to the following types of scenarios:

򐂰 Request/response: For example, a caller waits for a response. The caller
expects to receive a response quickly. There is no physical tragedy, such as a
transportation accident, that will occur if the response takes too long. Instead,
the caller might receive a time out, for example:

14 InfoSphere DataStage for Enterprise XML Data Integration

 – Business-to-consumer, such as web applications
– Online data entry applications
– Business-to-business
򐂰 The following DS solutions can be applied:
– ISD with SOAP or EJB binding
– ISD with JMS binding, when the caller expects the response to be placed
on a response queue

The term near-real time is used and applies to the following types of scenarios:
򐂰 Message delivery:
– The data is delivered and expected to be processed immediately.
– Users can accept a short lag time (ranging from seconds to a few
minutes).
– No person waits for a response, for example:
• Reporting systems
• Active warehouses
– The following DS solutions can be applied:
• MQ  DTS/UOW
• ISD with text over JMS binding

The high cost of starting up and shutting down jobs demanded that DataStage
was enhanced with additional capabilities to support these types of scenarios.
Implementing these types of scenarios with batch applications is not feasible.

1.4.3 Master Data Management

IBM InfoSphere Master Data Management (MDM) allows you to gain a unified,
complete, consistent and standardized view of your data to drive critical business
decisions. MDM also helps you gain control of your data, reduce information
errors and eliminate duplicate data. All of which helps you to meet your growth,
revenue-generation, and cost-reduction goals.

IBM InfoSphere MDM Server is a master repository that delivers a single version
of an organization’s data entities, such as customer, product, and supplier. Its
SOA library of prepackaged business services allows organizations to define
how they want users to access master data and seamlessly integrate into current
architectures and business processes.

Chapter 1. Heterogeneous data integration 15

 When business entities are deployed, typically data is loaded into an MDM
repository, but most data changes come from existing systems. MDM
implementations require a set of InfoSphere Information Server DataStage jobs
(see Figure 1-5) to perform the initial and delta loads directly to the MDM
database:
򐂰 An initial load is an original movement of data from source systems into the
MDM repository when the repository is empty.
򐂰 A delta load is a periodic (typically daily) data update from source systems
into MDM.

Source Systems MDM Server

Source Source So urce MDM Business

#1 #2 #N Information Server Serivces
User Interface
and
DataStage QS Duplicate Suspect Reporting
Processing
Information Server
Info rmation Analyzer SIF
Fast Track Load Process
DS Jobs MDM History
DataStag e Database

Figure 1-5 MDM server implementation

1.4.4 Web Services and XML

Web Services have gained prominence in the last few years. Web Services are
self-contained, modular, self-explanatory applications that are published on the
web. They provide functions to encapsulate objects from single functions to
complex business functions. Examples of simple functions are a calculator,
spreadsheet, or tutorial. Examples of complex functions are processing a tax
return, receiving stock quotes, or processing credit card transactions. After this
web service is deployed, anyone, either another application or web service, can
locate and invoke it.

Web Services involve a few components: SOAP; Universal Description,

Discovery, and Integration (UDDI); and Web Services Description Language
(WSDL). These components work together, as depicted in Figure 1-6 on
page 17, to interact between applications or clients, to publish themselves to be
consumed by other clients and to describe themselves.

16 InfoSphere DataStage for Enterprise XML Data Integration

 UDDI
UDDI
Registry
Registry
Another
Service Service
Broker Broker
Publish
with WSDL
Find through
WSDL

Service First Bind, then Invoke Service

Provider Requester

Web SOAP
Application
Service
by Client

Figure 1-6 Interaction of Web Services

SOAP is an XML-based protocol that allows applications to invoke applications

that are provided by service providers anywhere on the web. SOAP is supported
by HTTP and, therefore, can be run on the Internet without any new
requirements over an existing infrastructure. SOAP is independent of any
programming language and component technology, and it is object neutral. It is
also independent of operating systems.

Universal Description Discovery and Integration (UDDI) is a specification for

web registries of Web Services. Web users locate and discover services on a
UDDI-based registry. Registries of services are distributed over the Internet, and
these registries of services are described in a common XML format or schema.
With a common format, the searching and analysis of applications are easier. A
UDDI registry is available from the IBM alphaWorks® website, and it supports
users in various department-wide or company-wide scenarios.

Web Services Description Language (WSDL) is a language that is used to

describe a service to the world. WSDL is defined in the following description from
the World Wide Web Consortium (W3C)1:

“WSDL is an XML format for describing network services as a set of endpoints

operating on messages containing either document-oriented or
procedure-oriented information. The operations and messages are described
abstractly, and then bound to a concrete network protocol and message format to
define an endpoint. Related concrete endpoints are combined into abstract
endpoints (services). WSDL is extensible to allow the description of endpoints

1
World Wide Web Consortium at http://www.w3.org/TR/wsdl

Chapter 1. Heterogeneous data integration 17

 and their messages regardless of what message formats or network protocols
are used to communicate. However, the only bindings described in this
document describe how to use WSDL in conjunction with SOAP 1.1, HTTP
GET/POST, and MIME.”

1.4.5 Real-time integration using IBM InfoSphere Change Data

Capture
IBM InfoSphere Change Data Capture (CDC) is an enterprise data
synchronization and change capture solution. InfoSphere Change Data Capture
can be used to capture changes in near-real time as data is inserted, updated,
and deleted in tables in a source database. These captured changes are then
transformed to meet the target database characteristics and propagated to tables
in target databases, directly integrated into DataStage or trickle-fed to drive other
processes via messaging, as depicted in Figure 1-7.

The source and target databases can potentially be databases from separate
vendors running on separate platforms. Propagated data can involve user data
that is stored both in relational format, as well as XML, and can include audit data
that is hidden in the database logs.

Figure 1-7 InfoSphere Change Data Capture Enterprise Architecture

The flow of data through Change Data Capture can be divided into three parts.

18 InfoSphere DataStage for Enterprise XML Data Integration

Capture
Changes are captured by asynchronously reading from the database recovery
logs. Depending on the source database, InfoSphere Change Data Capture

might read the log files directly or it might use an application programming
interface (API) to read data from the log. InfoSphere Change Data Capture has
native log-based capture capabilities for the following databases:
򐂰 IBM DB2 z/OS®
򐂰 DB2 for IBM i
򐂰 DB2 Linux, UNIX, and Microsoft Windows
򐂰 Oracle
򐂰 SQL Server
򐂰 Sybase

Upon reading data from the database recovery logs, InfoSphere Change Data
Capture filters data based on the table where the change occurred. Only data
pertaining to tables of interest is retained for further processing. InfoSphere
Change Data Capture then stages the changes in a holding area until a commit
of these changes occurs in the source database. If a rollback occurs instead,
InfoSphere Change Data Capture discards the associated changes.

Transformations
In many cases, the data models of the source and target are not the same.
InfoSphere Change Data Capture can apply transformations to the data while it is
in flight between the source and target. The following transformations are
commonly used:
򐂰 Adding other information that can be obtained from the database recovery
logs
򐂰 Concatenation and other string functions
򐂰 Data type conversion
򐂰 Arithmetic
򐂰 Joining to look-up data in the secondary tables
򐂰 If/then/else logic

For more complex transformations, IBM InfoSphere DataStage is used in

conjunction with IBM InfoSphere Change Data Capture.

Apply
After transformations, the changes are applied to the target database. Changes
are applied by executing SQL statements against the appropriate target tables.

Chapter 1. Heterogeneous data integration 19

 InfoSphere Change Data Capture has native “Apply” capabilities for the following
databases:
򐂰 DB2 z/OS
򐂰 DB2 on IBM i
򐂰 DB2 Linux, UNIX, and Windows
򐂰 Oracle
򐂰 SQL Server
򐂰 Sybase
򐂰 Teradata

InfoSphere Change Data Capture also provides a level of customization around

the SQL statements that are executed as part of Apply processing. For example,
it is possible to perform these tasks:
򐂰 Ignore deletes that occur at the source
򐂰 Convert source updates and deletes into target inserts (an audit or change
history table)
򐂰 Convert source inserts into target updates (multiple source records merge to
one target record)

InfoSphere Change Data Capture also supports several non-database targets:

򐂰 Flat files
򐂰 InfoSphere DataStage
򐂰 WebSphere MQ
򐂰 JMS messaging
򐂰 Netezza®

1.4.6 Real-time integration using IBM Replication Server

InfoSphere Change Data Capture also works with InfoSphere Replication Server,
which is the IBM strategic data replication product for the IBM DB2 family.
InfoSphere Replication Server has a large well-established client base, is a key
component of the IBM Active-Active strategy, and is developed in close
cooperation with DB2 development teams. InfoSphere Replication Server
supports two replication architectures: SQL Replication that uses intermediate
staging tables and Q Replication that uses IBM WebSphere MQ for transport, as
shown in Figure 1-8 on page 21.

20 InfoSphere DataStage for Enterprise XML Data Integration

 S ource Ta rget
SOURCE2
CCD
SOURC E1
Capture
HISTORY
C D2
CD1
CD 1
User App REPLICA

UC OPY

DB Log

SQL Replication Architecture

Source Target
SOUR CE2
WebSphere MQ
SOURCE1

CC D

Capture User App UC OPY

DB Log

Q Replication Architecture

Figure 1-8 InfoSphere Replication Server Architecture

Similar to InfoSphere Change Data Capture, as described in 1.4.5, “Real-time

integration using IBM InfoSphere Change Data Capture” on page 18, InfoSphere
Replication Server also has Capture and Apply components, reads from DB2
logs, and performs data transformations prior to applying data to the target
database tables. The difference arises in the storage and transport methods that
are used, the supported data sources, and tight integration with DB2.

Propagated data can involve user data that is stored as relational data or as XML
and can include audit data from the database logs. Captured changes are also
transformed to meet the target database characteristics. Then, they are
propagated to target DB2 tables, directly integrated into DataStage, or trickle-fed
to drive other processes via messaging, as depicted in Figure 1-9 on page 22.

Chapter 1. Heterogeneous data integration 21

 Figure 1-9 Replication Server - DataStage Integration Architecture

1.4.7 Real-time integration using Data Event Publishers

IBM InfoSphere Data Event Publisher facilitates business integration by linking
changed data events in DB2 databases on Linux, UNIX, and Windows with
enterprise application integration (EAI) solutions. EAI solutions include
WebSphere Business Integrator, message brokers such as WebSphere
Message Broker, and data transformation tools such as InfoSphere DataStage.
DB2 changed-data events are captured and published as XML messages or
comma-separated value (CSV) data via WebSphere MQ that can be used by
other applications and tools to drive subsequent processing.

InfoSphere Data Event Publisher consists of the Capture component that is used
by InfoSphere Replication Server and WebSphere MQ transport. However, it
does not use the Apply component. Instead, the changed data is converted into
either XML messages or CSV data and written to the WebSphere MQ queue.
Consuming clients can retrieve the message at the other end of the queue and
do not provide any feedback to the Capture component. This type of architecture
eliminates the need for handshakes between a sending (Capture) process and a
receiving (Apply) process. Therefore, this architecture provides flexible
change-data solutions that enable many types of business integration scenarios,
as shown in Figure 1-10 on page 23.

22 InfoSphere DataStage for Enterprise XML Data Integration

 Sour ce Target
SOURCE2 WebSphere MQ
SOURCE1
U SER TAB LE
XML
Capture User App
C SV

User File
DB Log

Figure 1-10 Data Event Publisher Architecture

The following methods are examples of other integration approaches:

򐂰 Application-to-application integration: Changes to operational customer data
can be pushed to data integration solutions, such as InfoSphere DataStage
for updating customer relationship management (CRM) application data.
򐂰 Business process initiation: Adding a customer record can initiate a
WebSphere Business Integrator workflow that results in a welcome email, a
credit verification, and an update to the CRM system.
򐂰 Critical data event monitoring: Changed-data events can be delivered to
WebSphere MQ Broker so that a low-inventory level can be used to initiate a
product restocking workflow.
򐂰 Change-only data population: Data changes in operational systems can be
fed to InfoSphere DataStage or another extract, transform, and load (ETL)
product. The ETL product then populates a data warehouse, data mart, or
operational data store with the changes.

1.4.8 Real-time integration using Classic Data Event Publishers

Another type of Data Event Publisher is the IBM InfoSphere Classic Data Event
Publisher that drives business integration and ETL processes by publishing
changes made to IBM System z® data.

It captures database changes to IBM IMS™, VSAM, Computer Associates

CA-IDMS, or Software AG Adabas during active log journal processing or by
reading source-specific recovery logs. It therefore provides reliability and
recoverability that empower near-real-time delivery. Captured data is
automatically reformatted into a consistent relational format before it is published
to WebSphere MQ or Java Message Queues to simplify and maximize
integration possibilities.

InfoSphere Classic Data Event Publisher for z/OS V9.5 provides a new
publication interface for use by InfoSphere DataStage and InfoSphere DataStage

Chapter 1. Heterogeneous data integration 23

 for Linux on System z. Data received from existing sources can be processed
and converted into XML using the XML Transformation stage within InfoSphere
DataStage.

1.4.9 Summary
In this chapter, we discuss the benefits of using XML technologies that are widely
accepted by various industrial organizations throughout the world. We also
describe a number of IBM solutions that facilitate data integration and how to use
current and future investments. The focus of this book is the XML integration
capabilities that are built into DataStage. It is imperative to have a basic
understanding of XML and a working knowledge of DataStage to fully appreciate
the concepts and solutions that are presented in this book.

In subsequent chapters, we provide an overview of XML and an introduction to

DataStage. The following chapters provide detailed information about the XML
capabilities available in DataStage, working with the XML Pack and the new XML
Transformation stage, and performance considerations for working with large and
complex XML data. The final chapter provides examples of integration scenarios
and use case scenarios.

24 InfoSphere DataStage for Enterprise XML Data Integration

 2

Chapter 2. XML overview

In this chapter, we introduce you to the eXtensible Markup Language (XML). This
book is not designed to teach you everything about XML, but it is important to
understand the basic concepts. In the following sections, we describe what XML
is (and is not), and discuss parsers, schemas, namespaces, Extensible
Stylesheet Language (XSL), Web services, and other XML-related technologies.

© Copyright IBM Corp. 2012. All rights reserved. 25

2.1 XML introduction
The idea of universal data formats is not new. Programmers tried to find ways to
exchange information between computer programs for a long time. First,
Standard Generalized Markup Language (SGML) was developed. SGML can be
used to mark up data, that is, to add metadata in a way that allows data to be
self-describing. SGML is metalanguage.

Extensible Markup Language (XML) is a simple, flexible text format that is

designed to be used in data interchange on the web and in other electronic
communications.

XML was developed in 1998 and is now widely used. It is one of the most flexible
ways to automate web transactions. XML is derived as a subset from Standard
Generalized Markup Language (SGML) and is designed to be simple, concise,
human readable, and relatively easy to use in programs on multiple platforms.
For more information about the XML standard, see the following web page:
http://www.w3.org/XML

As with other markup languages, XML is built using tags. Basic XML consists of
start tags, end tags, and a data value between the two. In XML, you create your
own tags, with a few restrictions. Example 2-1 shows a simple XML document.

Example 2-1 Simple XML document

<?xml version="1.0" encoding="UTF-8" ?>
<Employee>
<ID>0000002150</ID>
<Lastname>SMITH</Lastname>
<Firstname>ROBERT</Firstname>
<Company>IBM</Company>
<Address>Bytoften 1</Address>
<Zipcode>8240</Zipcode>
<City>Risskov</City>
<Country>Denmark</Country>
</Employee>

The XML syntax is simple, but it is difficult to parse and transform an XML
document into a form that is usable to programming languages. Therefore, it is
essential to have access to efficient parsing and transformation tools.

XML contains document type and schema definitions. These document type and
schema definitions are used to specify semantics (allowable grammar) for an
XML document.

26 InfoSphere DataStage for Enterprise XML Data Integration

2.2 The value of XML data
As a result of XML industry standards becoming more prevalent, the drive toward
service-oriented architecture (SOA) environments, and rapid adoption of
syndication technologies, more XML data is generated every day. XML is in web
feeds, purchase orders, transaction records, messages in SOA environments,
financial trades, insurance applications, and other industry-specific and
cross-industry data. That is, XML data and documents are becoming important
business assets that contain valuable information, such as customer details,
transaction data, web logs, and operational documents.

The growth and pervasiveness of XML assets present challenges and

opportunities for the enterprise. When XML data is harnessed, and the value of
the information it contains is unlocked, it can translate into opportunities for
organizations to streamline operations, derive insight, and become agile.

Alternatively, as XML data becomes more critical to the operations of an

enterprise, it presents challenges in that XML data must be secured, maintained,
searched, and shared. Depending on its use, XML data might also need to be
updated, audited, and integrated with traditional data. All of these tasks must be
performed with the reliability, availability, and scalability afforded to traditional
data assets.

Unleashing the potential of XML data requires a robust information integration

platform. In the next section, we provide several scenarios that drive the use of
XML technologies.

SOA and Web services

Services-based frameworks and deployments can more easily be used to
integrate systems, reuse resources, and enable fast reactions to changing
market conditions. In services-based architectures, consumers and service
providers exchange information using messages. These messages are invariably
encapsulated as XML, which can provide the foundation in SOA environments,
as depicted in Figure 2-1. Thus, the drive toward information as a service and the
adoption of SOA environments is also stimulating the growth of XML.

XML
Service Service
Requester Provider

Figure 2-1 XML: The foundation for Web services

Chapter 2. XML overview 27

 Web 2.0 technologies
Syndication is considered to be the heartbeat of Web 2.0, the next generation of
the Internet. Atom and RSS feeds exist in abundance on the web, allowing you to
subscribe to them and to be kept up-to-date about all kinds of web content
changes, such as news stories, articles, wikis, audio files, and video files.
Content for these feeds is rendered as XML files and can contain links,
summaries, full articles, and even attached multimedia files, such as podcasts.
Syndication and web feeds are transforming the web as we know it. New
business models are emerging around these technologies. As a consequence,
XML data now exists not only in companies adopting XML industry standards, or
enterprises implementing SOAs, but also on virtually every web-connected
desktop.

2.3 Introduction to XML documents

XML is reminiscent of HTML, because they are both derived from SGML, which
was defined in 1986. But unlike HTML, XML tags identify the data, rather than
specifying how to display it. For example, an HTML tag might display this data in
bold font (<b>...</b>). An XML tag acts like a field name in your program. It puts
a label on a piece of data to identify it (<message>...</message>).

XML documents can be well formed, or they can be well formed and valid. These
important rules do not exist for HTML documents. These rules contrast with the
freestyle nature of many of the concepts in XML. The rules can be defined briefly
as invalid, valid, and well-formed documents.

There are three kinds of XML documents:

򐂰 Invalid documents do not follow the syntax rules defined by the XML
specification. If a developer defined rules for the document content in a
document type definition (DTD) or schema, and the document does not follow
those rules, that document is invalid as well. (See 2.4, “Defining document
content” on page 32, for a proper introduction to DTDs and schemas for XML
documents.)
򐂰 Valid documents follow both the XML syntax rules and the rules defined in
their DTD or schema.
򐂰 Well-formed documents follow the XML syntax rules but do not have a DTD or
schema.

28 InfoSphere DataStage for Enterprise XML Data Integration

2.3.1 Well-formed XML documents
To be well-formed documents, the documents must conform to the basic rules for
XML documents. In Table 2-1, we provide a list of a subset of the requirements
for well-formed XML documents.

Table 2-1 Requirements for a well-formed document

Requirement Well-formed example Not well-formed example

The document has <element0> <element1>value1</element1>

exactly one root <elem1>value1</elem1> <element2>value2</element2>
element. <elem2>value2</elem2>
</element0>

Each start tag is <elem1> <elem1>

matched by one end <elem2>value2</elem2> <elem2>value2
tag. </elem1> </elem1>

All elements are <elem1> <elem1>

properly nested. <elem2>value2</elem2> <elem2>value2</elem1>
</elem1> </elem2>

Attribute values are <elem1 <elem1 id=15>value1</elem1>

quoted. id=”15”>value1</elem1>

XML elements are <elem1> <ELEM1>

case-sensitive. <elem2>value2</elem2> <elem2>value2</ELEM2>
</elem1> </elem1>

Disallowed characters <elem1> <elem1> 123<456</elem1>

are not used in tags or 123&lt;456</elem1>
values.

Validation
The process of checking to see whether an XML document conforms to a
schema or DTD is called validation. Validation is in addition to checking a
document for compliance to the XML core concept of syntactic well-formedness.
All XML documents must be well formed, but it is not required that a document is
valid unless the XML parser is validating. When validating, the document is also
checked for conformance with its associated schema.

Documents are only considered valid if they satisfy the requirements of the DTD
or schema with which they are associated. These requirements typically include
the following types of constraints:
򐂰 Elements and attributes that must or might be included, and their permitted
structure.

Chapter 2. XML overview 29

 򐂰 The structure is specified by a regular expression syntax.

InfoSphere Information Server DataStage provides robust XML parsing and

validation capabilities.

XML declarations
Most XML documents start with an XML declaration that provides basic
information about the document to the parser. An XML declaration is
recommended, but not required. If there is an XML declaration, it must be the first
item in the document.

The declaration can contain up to three name-value pairs (many people call them
attributes, although technically they are not). The version is the version of XML
used; currently this value must be 1.0. The encoding is the character set used in
this document. The ISO-8859-1 character set referenced in this declaration
includes all of the characters used by most Western European languages. If no
encoding is specified, the XML parser assumes that the characters are in the
UTF-8 set, a Unicode standard that supports virtually every character and
ideograph in all languages. The following declaration is an example:
<?xml version="1.0" encoding="ISO-8859-1" standalone="no"?>

Finally, standalone, which can be either yes or no, defines whether this
document can be processed without reading any other files. For example, if the
XML document does not reference any other files, you specify standalone="yes".
If the XML document references other files that describe what the document can
contain (more about those files later), you specify standalone="no". Because
standalone="no" is the default, you rarely see standalone in XML declarations.

Namespaces
The power of XML comes from its flexibility. You can define your own tags to
describe your data. Consider an XML document with the tag <Title>. The
document includes the <title> element for a courtesy title, the <title> element for
the title of a book, and the <title> element for the title to a piece of property. All
the <Title> information elements mentioned in this case are reasonable choices,
but all of them create elements with the same name. How do you tell if a specific
<title> element refers to a person, a book, or a piece of property? These
situations are handled with namespaces.

To use a namespace, define a namespace prefix and map it to a particular string.

In Example 2-2, we show how you might define namespace prefixes for the three
<title> elements.

Example 2-2 Define namespace prefixes

<?xml version="1.0"?>

30 InfoSphere DataStage for Enterprise XML Data Integration

<customer_summary
xmlns:addr="http://www.xyz.com/addresses/"
xmlns:books="http://www.zyx.com/books/"
xmlns:mortgage="http://www.yyz.com/title/"
>
... <addr:name><title>Mrs.</title> ... </addr:name> ...
... <books:title>Lord of the Rings</books:title> ...
... <mortgage:title>NC2948-388-1983</mortgage:title> ...

In Example 2-2 on page 30, the three namespace prefixes are addr, books, and
mortgage. Defining a namespace for a particular element means that all of its
child elements belong to the same namespace. The first <title> element belongs
to the addr namespace because its parent element, <addr:Name>, belongs to
the addr namespace.

Additional components
Comments, processing instructions, and entities are additional components of
XML. The following sections provide additional details about these components.

Comments
Comments can appear anywhere in the document; they can even appear before
or after the root element. A comment begins with <!-- and ends with -->. A
comment cannot contain a double hyphen (--) except at the end; with that
exception, a comment can contain anything. Most importantly, any markup inside
a comment is ignored; if you want to remove a large section of an XML
document, simply wrap that section in a comment. (To restore the
commented-out section, simply remove the comment tags.) This example
markup contains a comment:
<!-- This is comment section of Customer Name XML Doc : -->
<Customer Name>
<First Name> John </First Name>
<Last Name> Smith </Last Name>
</Customer Name>

Processing instructions
A processing instruction is markup intended for a particular piece of code. In the
following example, a processing instruction (PI) exists for Cocoon, which is an
XML processing framework from the Apache Software Foundation. When
Cocoon processes an XML document, it looks for processing instructions that
begin with cocoon-process, then processes the XML document. In this example,
the type="sql" attribute tells Cocoon that the XML document contains an SQL
statement:
<!-- Here is a PI for Cocoon: -->

Chapter 2. XML overview 31

 <?cocoon-process type="sql"?>

Entities
The following example defines an entity for the document. Anywhere the XML
processor finds the string &RB;, it replaces the entity with the string ‘RedBooks’:
<!-- Here is an entity: -->
<!ENTITY RB "RedBooks">

The XML spec also defines five entities that can be used in place of various
special characters:
򐂰 &lt; for the less-than sign
򐂰 &gt; for the greater-than sign
򐂰 &quot; for a double-quote
򐂰 &apos; for a single quote (or apostrophe)
򐂰 &amp; for an ampersand

2.4 Defining document content

Document type definitions and XML schema are both used to describe structured
information. XML schema is gaining acceptance. Both document type definitions
and schemas are building blocks for XML documents and consist of elements,
tags, attributes, and entities.

2.4.1 Document type definition

A document type definition (DTD) specifies the kinds of tags that can be included
in your XML document, the valid arrangements of those tags, and the structure of
the XML document. The DTD defines the type of elements, attributes, and
entities allowed in the documents, and it can also specify limitations to their
arrangement. You can use the DTD to ensure that you do not create an invalid
XML structure. The DTD defines how elements relate to one another within the
document tree structure. You can also use it to define which attributes can be
used to define an element and which attributes are not allowed.

A DTD defines your own language for a specific application. The DTD can be
either stored in a separate file or embedded within the same XML file. If it is
stored in a separate file, it can be shared with other documents. XML documents
referencing a DTD contain a <!DOCTYPE> declaration. This <!DOCTYPE>
declaration either contains the entire DTD declaration in an internal DTD or
specifies the location of an external DTD. Example 2-3 on page 33 shows an
external DTD in a file named DTD-Agenda.dtd.

32 InfoSphere DataStage for Enterprise XML Data Integration

Example 2-3 An external DTD
<?xml version="1.0" encoding="UTF-8"?>
<!ELEMENT AGENDA (PERSONA+)>
<!ELEMENT PERSONA (EMPRESA, CARGO, NOMBRE, TELEFONO1+,TELEFONO2*,
EXT?)>
<!ELEMENT EMPRESA (#PCDATA)>
<!ELEMENT CARGO (#PCDATA)>
<!ELEMENT NOMBRE (#PCDATA)>
<!ELEMENT TELEFONO1 (#PCDATA)>
<!ELEMENT TELEFONO2 (#PCDATA)>
<!ELEMENT EXT (#PCDATA)>

Example 2-4 is an XML document that refers to this external DTD.

Example 2-4 Reference to an external DTD

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE AGENDA SYSTEM "DTD_Agenda.dtd">
<AGENDA>
<PERSONA>
<EMPRESA>Matutano</EMPRESA>
<CARGO>Gerente</CARGO>
<NOMBRE>Pepe Montilla</NOMBRE>
<TELEFONO1>912563652</TELEFONO1>
<TELEFONO2>658968574</TELEFONO2>
<EXT>256</EXT>
</PERSONA>
</AGENDA>

We can also define an internal DTD in an XML document so that both of them
are in the same file, as shown in Example 2-5. In either case, internal or external
DTD, the <!DOCTYPE> declaration indicates the root element.

Example 2-5 An internal DTD

<!DOCTYPE RAIZ [
<!ELEMENT ...... >
<!ELEMENT ...... >
<!ELEMENT ...... >
<!ELEMENT ...... >
]>
<RAIZ>
.
..

Chapter 2. XML overview 33

 </RAIZ>

A DTD is not required in an XML document. However, a DTD provides parsers

with clear instructions about what to check when they are determining the validity
of an XML document. DTDs or other mechanisms, such as XML schemas,
contribute to the goal of ensuring that the application can easily determine
whether the XML document adheres to a specific set of rules, beyond the
well-formedness rules that are defined in the XML standard. DTDs are also used
by tools to create XML documents.

Symbols in DTDs
A few symbols are used in DTDs to indicate how often (or whether) something
can appear in an XML document. We show several examples, along with their
meanings:
򐂰 <!ELEMENT address (name, city, state)>
The <address> element must contain a <name>, a <city>, and a <state>
element, in that order. All of the elements are required. The comma indicates
a list of items.
򐂰 <!ELEMENT name (title?, first-name, last-name)>
This means that the <name> element contains an optional <title> element,
followed by a mandatory <first-name> and a <last-name> element. The
question mark indicates that an item is optional; it can appear one time or not
at all.
򐂰 <!ELEMENT addressbook (address+)>
An <addressbook> element contains one or more <address> elements. You
can have as many <address> elements as you need, but there must be at
least one. The plus sign (+) indicates that an item must appear at least one
time, but it can appear any number of times.
򐂰 <!ELEMENT private-addresses (address*)>
A <private-addresses> element contains zero or more <address> elements.
The asterisk indicates that an item can appear any number of times, including
zero.
򐂰 <!ELEMENT name (title?, first-name, (middle-initial | middle-name)?,
last-name)>
A <name> element contains an optional <title> element, followed by a
<first-name> element, possibly followed by either a <middle-initial> or a
<middle-name> element, followed by a <last-name> element. Both
<middle-initial> and <middle-name> are optional, and you can have only one
of the two elements. Vertical bars indicate a list of choices; you can choose

34 InfoSphere DataStage for Enterprise XML Data Integration

 only one item from the list. This example uses parentheses to group certain
elements, and it uses a question mark against the group.
򐂰 <!ELEMENT name ((title?, first-name, last-name) | (surname, mothers-name,
given-name))>
The <name> element can contain one of two sequences: An optional <title>,
followed by a <first-name> and a <last-name>; or a <surname>, a
<mothers-name>, and a <given-name>.

2.4.2 XML schema

The World Wide Web Consortium (W3C) XML Schema Definition Language
(XSDL) is an XML language for describing and constraining the content of XML
documents. A schema is similar to a DTD in that it defines which elements an
XML document can contain, how they are organized, and which attributes and
attribute types elements can be assigned. The Information Server Schema
Manager does not support DTDs, but the Information Server supports XML
schemas. Therefore, a schema is a method to check the validity of well-formed
XML documents. The following list shows the major advantages of schemas over
DTDs:
򐂰 XML schemas use XML syntax.
An XML schema is an XML document. You can process a schema just like
any other document. For example, you can write an XSLT style sheet that
converts an XML schema into a web form complete with automatically
generated JavaScript code that validates the data as you enter it.
򐂰 XML schemas support data types.
DTDs support data types; however, it is clear those data types were
developed from a publishing perspective. XML schemas support all of the
original data types from DTDs, such as IDs and ID references. They also
support integers, floating point numbers, dates, times, strings, URLs, and
other data types that are useful for data processing and validation.
򐂰 XML schemas are extensible.
In addition to the data types defined in the XML schema specification, you can
also create your own, and you can derive new data types based on other data
types.
򐂰 XML schemas have more expressive power.
For example, with XML schemas, you can define that the value of any <state>
attribute cannot be longer than two characters, or that the value of any
<postal-code> element must match the regular expression [0-9]{5}(-[0-9]{4})?.
You cannot perform either of these functions with DTDs.

Chapter 2. XML overview 35

 Sample XML schema
It is difficult to give a general outline of the elements of a schema due to the
number of elements that can be used according to the W3C XML Schema
Definition Language. The purpose of this language is to provide an inventory of
XML markup constructs with which to write schemas. Example 2-6 is a simple
document that describes the information about a book.

Example 2-6 A book description

<?xml version="1.0" encoding="UTF-8"?>
<book isbn="0836217462">
<title>
Don Quijote de la Mancha
</title>
<author>De Cervantes Saavedra, Miguel</author>
<character>
<name>Sancho Panza</name>
<friend-of>El Quijote</friend-of>
<since>1547-10-04</since>
<qualification> escudero </qualification>
</character>
<character>
<name>ElbaBeuno</name>
<since>1547-08-22</since>
<qualification>Amor Platonico de Don Quijote</qualification>
</character>
</book>

Because the XML schema is a language, several choices exist to build a possible
schema that covers the XML document. Example 2-7 is a possible, simple
design.

Example 2-7 XML schema

<?xml version="1.0" encoding="utf-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="book">
<xs:complexType>
<xs:sequence>
<xs:element name="title" type="xs:string"/>
<xs:element name="author" type="xs:string"/>
<xs:element name="character" minOccurs="0" maxOccurs="unbounded">
<xs:complexType>
<xs:sequence>
<xs:element name="name" type="xs:string"/>
<xs:element name="friend-of" type="xs:string" minOccurs="0"

36 InfoSphere DataStage for Enterprise XML Data Integration

 maxOccurs="unbounded"/>
<xs:element name="since" type="xs:date"/>
<xs:element name="qualification" type="xs:string"/>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:sequence>
<xs:attribute name="isbn" type="xs:string"/>
</xs:complexType>
</xs:element>
</xs:schema>

Example 2-7 on page 36 is obviously an XML document, because it begins with

the XML document declaration. The schema element opens our schema holding
the definition of the target namespace. Then, we define an element named book.
This element is the root element in the XML document. We decided that it is a
complex type, because it has attributes and non-text children. With sequence, we
begin to declare the children elements of the root element book. W3C XML
schema lets us define the type of data, as well as the number of possible
occurrences of an element. To learn about the possible values for these types,
see the specification documents from W3C.

XML schema options

The XML schema language offers possibilities and alternatives beyond what is
shown in Example 2-7 on page 36. We can develop another schema based on a
flat catalog of all the elements available in the instance document and, for each
of the elements, lists of child elements and attributes. Thus, we have the
following two choices:
򐂰 Defining elements and attributes as they are needed
򐂰 Creating the elements first and referencing them

The first option has a real disadvantage: the schema might be difficult to read
and maintain when documents are complex. W3C XML schema allows us to
define data types and use these types to define our attributes and elements. It
also allows the definition of groups of elements and attributes. In addition, there
are several ways to arrange relationships between the elements.

2.4.3 Namespaces
Namespaces are used when you need separate elements, possibly with separate
attributes, but with the same name. Depending upon the context, a tag is related
to one specific element or to another specific element. Without namespaces, it is

Chapter 2. XML overview 37

 impossible to specify the context to which an element belongs. Example 2-8
shows this situation.

Example 2-8 XML document with duplicated element names

<?xml version="1.0" ?>
<book>
<title>XML Sample</title>
<pages>210</pages>
<isbn>1-868640-34-2</isbn>
<author>
<firstname>JuanJose</firstname>
<lastname>Hernandez</lastname>
<title>Mr</title>
</author>
</book>

Clearly, a problem exists with the element <title>. It appears in two separate
contexts. This situation complicates the situation for processors and might cause
ambiguities. We need a mechanism to distinguish between the two contexts and
apply the correct semantic description to each tag. The cause of this problem is
that this document uses only one common namespace. The solution to this
problem is namespaces. Namespaces are a simple and straightforward way to
distinguish names that are used in XML documents. By providing the related
namespace when an element is validated, the problem is solved, as shown in
Example 2-9.

Example 2-9 XML document using namespace prefixes

<?xml version="1.0" ?>
<library-entry
xmlns:authr="http://sc58ts.itso.ibm.com/Jose/2002/author.dtd"
xmlns:bk="books.dtd">
<bk:book>
<bk:title>XML Sample</bk:title>
<bk:pages>210</bk:pages>
<bk:isbn>1-868640-34-2</bk:isbn>
<authr:author>
<authr:firstname>JuanJose</authr:firstname>
<authr:lastname>Hernandez</authr:lastname>
<authr:title>Mr</authr:title>
</authr:author>
</bk:book>

38 InfoSphere DataStage for Enterprise XML Data Integration

 As you can see in Example 1-8, the <title> tag is used twice, but in separate
contexts: within the <author> element and within the <book> element. Note the
use of the xmlns keyword in the namespace declaration. After a prefix is defined
on an element, it can then be used by all descendants of that element. In
Example 2-10 on page 41, we specified the relevant namespace prefix before
each element to illustrate the relationship of each element to a specific
namespace. However, if the prefix is not specified, the element is in the default
namespace if a default namespace is specified or not in a namespace if there is
no default namespace. A default namespace is defined on an element with the
xmlns attribute without specifying a prefix.

2.5 Extensible Stylesheet Language

Up to this point, we described XML, its syntax, how XML is used to mark up
information according to our own vocabularies, and how a program can check the
validity of an XML document. We described how XML can ensure that an
application running on any particular platform receives valid data. XML helps you
ensure that a program can process this data. However, because XML describes
document syntax only, the program does not know how to format this data
without specific instructions about style.

The solution is to use XSL transformations. The Extensible Stylesheet Language

(XSL) specification describes powerful tools to accomplish the required
transformation of XML data:
򐂰 The XSL Transformations (XSLT) language is for transformation.
򐂰 Formatting Objects (FO) is a vocabulary for describing the layout of
documents.
򐂰 XSLT uses the XML Path Language (XPath), which is a separate
specification that describes a means of addressing XML documents and
defining simple queries.

XSLT offers a powerful means of transforming XML documents into other forms,
producing XML, HTML, and other formats. It can sort, select, and number. And, it
offers many other features for transforming XML. It operates by reading a style
sheet, which consists of one or more templates, and then matching the templates
as it visits the nodes of the XML document. The templates can be based on
names and patterns. In the context of information integration, XSLT is
increasingly used to transform XML data into another form, sometimes other
XML (for example, filtering out certain data, SQL statements, and plain text), or
any other format. Thus, any XML document can be shown in separate formats,
such as HTML, PDF, RTF, VRML, and Postscript.

Chapter 2. XML overview 39

2.5.1 XPath
XPath is a string syntax for building addresses to the information that is found in
an XML document. We use this language to specify the locations of document
structures or data found in an XML document when processing that information
using XSLT. XPath allows us from any location to address any other location or
content. XPath is a tool that is used in XSLT to select certain information to be
formatted.

XPath patterns
Table 2-2 shows several XPath patterns. These examples show you a few objects
that you can select.

Table 2-2 XPath pattern

Symbol Meaning

/ Refer to the immediate child

// Refer to any child in the node

. Refer to actual context

* Refer to all elements in the actual node

@ Refer to an attribute

@* Refer to all attributes in the actual node

XPath models an XML document as a tree of nodes, as shown in the following

list:
򐂰 Root nodes
򐂰 Element nodes
򐂰 Attribute nodes
򐂰 Text nodes
򐂰 Namespace nodes
򐂰 Processing instruction nodes
򐂰 Comment nodes

The basic syntactic construct in XPath is the expression, as shown in

Example 2-10 on page 41. An object is obtained by evaluating an expression,
which has one of the following four basic types:
򐂰 Node-set (an unordered collection of nodes without duplicates)
򐂰 Boolean
򐂰 Number
򐂰 String

40 InfoSphere DataStage for Enterprise XML Data Integration

 Example 2-10 Xpath
<?xml version="1.0"?>
<!DOCTYPE library system "library.dtd">
<library>
<book ID="B1.1">
<title>xml</title>
<copies>5</copies>
</book>
<book ID="B2.1">
<title>WebSphere</title>
<copies>10</copies>
</book>
<book ID="B3.2">
<title>great novel</title>
<copies>10</copies>
</book>
<book ID="B5.5">
<title>good story</title>
<copies>10</copies>
</book>
</library>

Considering Example 2-10, we can make the following paths:

򐂰 /book/copies selects all copies of element children of book elements.
򐂰 /book//title selects all title elements in the tree, although title elements are not
immediate children.

The /book/@ID path selects all ID attributes beyond book elements.

It is also possible to select elements based on other criteria. The

/library/*/book[title $eq$ "good story"] path selects all book elements beyond the
library element, but only if the title element matches “good story”.

2.5.2 XSLT
XSLT helps you access and display the content in the XML file. XSLT is referred
to as the stylesheet language of XML. The relationship of Cascading Style
Sheets (CSS) and HTML is comparable to the relationship of XSLT and XML.
However, XML and XSLT are far more sophisticated technologies than HTML
and CSS.

XSLT is a high-level declarative language. It is also a transforming and formatting

language. It behaves in the following way:

Chapter 2. XML overview 41

 򐂰 The pertinent data is extracted from an XML source document and
transformed into a new data structure that reflects the desired output. The
XSLT markup is commonly called a stylesheet. A parser is used to convert the
XML document into a tree structure that is composed of various types of
nodes. The transformation is accomplished with XSLT by using pattern
matching and templates. Patterns are matched against the source tree
structure, and templates are used to create a result tree.
򐂰 Next, the new data structure is formatted, for example, in HTML or as text,
and finally the data is ready for display.

Figure 2-2 shows the source tree from the XML document that was presented in
Example 2-10 on page 41.

library

book book book book

title title title title

copies copies copies copies

Figure 2-2 Source tree

The resultant tree after an XSL transformation might be an XHTML document, as

shown in Figure 2-3.

XHTML

HEAD BODY

title P

book

title

copies

Figure 2-3 Source tree after an XML transformation

42 InfoSphere DataStage for Enterprise XML Data Integration

 Based on how we instruct the XSLT processor to access the source of the data
that is transformed, the processor incrementally builds the result by adding the
filled-in templates. We write our stylesheets, or “transformation specifications,”
primarily with declarative constructs, although we can employ procedural
techniques if and when needed. We assert the desired behavior of the XSLT
processor based on conditions found in our source. XSLT manipulates the
source tree only, and the original XML document is left unchanged.

The most important aspect of XSLT is that it allows you to perform complex
manipulations on the selected tree nodes by affecting both content and
appearance. The final output might not resemble the source document. XSLT far
surpasses CSS in this capability to manipulate the nodes.

2.6 Web services

Web services offer a new dimension to information integration. Consider Web
services as a gateway to service-oriented architecture (SOA) for the data
integration needs of an enterprise. You can use Web services to extract and
integrate business data from heterogeneous business systems. By providing
interface standards, Web services can be an important component of information
as a service architecture.

Web services allow applications to communicate with each other in a

platform-independent and programming language-independent manner. A web
service is a software interface that describes a collection of operations that can
be accessed over the network through standardized XML messaging. A web
service uses protocols based on the XML language to describe an operation to
execute or data to exchange with another web service. A group of Web services
interacting together in this manner defines a particular web service application in
an SOA.

Web services help to bridge the gap between business users and technologists
in an organization. Web services make it easier for business users to understand
web operations. Business users can then describe events and activities, and
technologists can associate them with appropriate services.

With universally defined interfaces and well-designed tasks, it also is easier to

reuse these tasks and the applications that they represent. Reusability of
application software means a better return on investment on software, because it
can produce more from the same resources. Reusability of application software
allows business people to consider using an existing application in a new way or
offering it to a partner in a new way, thus potentially increasing the business
transactions between partners.

Chapter 2. XML overview 43

 Therefore, Web services address the issues of data and application integration
and of transforming technical functions into business-oriented computing tasks.
These two facets allow businesses to communicate on a process level or
application level with their partners, while leaving dynamic room to adapt to new
situations or work with different partners on demand.

The versatility of XML is what makes Web services different from previous
generation component technologies. XML allows the separation of grammatical
structure (syntax) and the grammatical meaning (semantics), and how that is
processed and understood by each service and the environment in which it
exists. Now, objects can be defined as services, communicating with other
services in XML-defined grammar, whereby each service then translates and
analyzes the message according to its local implementation and environment.
Thus, a networked application can be composed of multiple entities of various
makes and designs as long as they conform to the rules defined by their SOA.

Web services allow you to perform the following tasks:

򐂰 Interact between services on any platform, written in any language.
򐂰 Conceptualize application functions into tasks, leading to task-oriented
development and workflows. This conceptualization allows a higher
abstraction of software that can be employed by less software-technical users
that work on business-level analytics.
򐂰 Allow for loose-coupling, which means that interactions between service
applications might not break each time there is a change in how services are
designed or implemented.
򐂰 Adapt existing applications to changing business conditions and customer
needs.
򐂰 Provide existing software applications with service interfaces without
changing the original applications, allowing them to fully operate in the
service environment.
򐂰 Introduce other administrative or operations management functions, such as
reliability, accountability, and security, independently of the original function,
thus increasing its versatility and usefulness in the business computing
environment.

44 InfoSphere DataStage for Enterprise XML Data Integration

 3

Chapter 3. Overview of IBM InfoSphere

DataStage
You can take advantage of the capabilities of InfoSphere DataStage when you
understand the underlying architecture. In this chapter, we describe InfoSphere
Information Server (IS) architecture, InfoSphere DataStage, and the
corresponding XML transformation architecture. Additionally, we explain the
DataStage elements that are used in integration scenarios from an architectural
perspective.

© Copyright IBM Corp. 2012. All rights reserved. 45

3.1 Introduction to IBM InfoSphere Information Server
Over the years, most organizations made significant investments in enterprise
resource planning, customer relationship management, and supply chain
management packages in addition to their home-grown applications. These
investments resulted in larger amounts of data captured about their businesses.
To turn all this data into consistent, timely, and accurate information for
decision-making requires an effective means of integrating information.

IBM InfoSphere Information Server is the first comprehensive, unified foundation

for enterprise information architectures, capable of scaling to meet any
information volume requirement so that companies can deliver business results
within these initiatives faster and with higher quality results. InfoSphere
Information Server provides a single unified platform that enables companies to
understand, cleanse, transform, and deliver trustworthy and context-rich
information.

InfoSphere Information Server helps you derive more value from complex,
heterogeneous information. It helps business and IT personnel collaborate to
understand the meaning, structure, and content of information across a wide
variety of sources. IBM InfoSphere Information Server helps you access and use
information in new ways to drive innovation, increase operational efficiency, and
lower risk.

Most critical business initiatives cannot succeed without the effective integration
of information. Initiatives, such as a single view of the customer, business
intelligence, supply chain management, and Basel II and Sarbanes-Oxley
compliance, require consistent, complete, and trustworthy information.

InfoSphere Information Server supports all of these initiatives:

򐂰 Business intelligence: InfoSphere Information Server makes it easier develop
a unified view of the business for better decisions. It helps you understand
existing data sources, cleanse, correct, and standardize information, and load
analytical views that can be reused throughout the enterprise.
򐂰 Master data management: InfoSphere Information Server simplifies the
development of authoritative master data by showing where and how
information is stored across source systems. It also consolidates disparate
data into a single, reliable record, cleanses and standardizes information,
removes duplicates, and links records across systems. This master record
can be loaded into operational data stores, data warehouses, or master data
applications, such as IBM WebSphere Customer Center. The record can also
be assembled, completely or partially, on demand.

46 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 Infrastructure rationalization: InfoSphere Information Server aids in reducing
operating costs by showing relationships among systems and by defining
migration rules to consolidate instances or move data from obsolete systems.
Data cleansing and matching ensure high-quality data in the new system.
򐂰 Business transformation: InfoSphere Information Server can speed
development and increase business agility by providing reusable information
services that can be plugged into applications, business processes, and
portals. These standards-based information services are maintained centrally
by information specialists but are widely accessible throughout the enterprise.
򐂰 Risk and compliance: InfoSphere Information Server helps improve visibility
and data governance by enabling complete, authoritative views of information
with proof of lineage and quality. These views can be made widely available
and reusable as shared services, while the rules inherent in them are
maintained centrally.

3.1.1 Capabilities
InfoSphere Information Server features a unified set of separately orderable
product modules, or suite components, that solve multiple types of business
problems. Information validation, access, and processing rules can be reused
across projects, leading to a higher degree of consistency, stronger control over
data, and improved efficiency in IT projects. This single unified platform enables
companies to understand, cleanse, transform, and deliver trustworthy and
context-rich information, as shown in Figure 3-1 on page 48.

Chapter 3. Overview of IBM InfoSphere DataStage 47

Figure 3-1 InfoSphere Information Server Architecture

Understand the data

InfoSphere Information Server can help you automatically discover, define, and
model information content and structure, and understand and analyze the
meaning, relationships, and lineage of information. By automating data profiling
and data-quality auditing within systems, organizations can achieve these goals:
򐂰 Understand data sources and relationships
򐂰 Help eliminate the risk of using or proliferating bad data
򐂰 Improve productivity through automation
򐂰 Use existing IT investments

InfoSphere Information Server makes it easier for businesses to collaborate

across roles. Data analysts can use analysis and reporting functionality,
generating integration specifications and business rules that they can monitor
over time. Subject matter experts can use web-based tools to define, annotate,
and report on fields of business data. A common metadata foundation enables

48 InfoSphere DataStage for Enterprise XML Data Integration

different types of users to create and manage metadata by using tools that are
optimized for their roles.

Cleanse the information

InfoSphere Information Server supports information quality and consistency by
standardizing, validating, matching, and merging data. It can certify and enrich
common data elements, use trusted data such as postal records for name and
address information, and match records across or within data sources.
InfoSphere Information Server allows a single record to survive from the best
information across sources for each unique entity, helping you to create a single,
comprehensive, and accurate view of information across source systems.

Transform the data into information

InfoSphere Information Server transforms and enriches information to ensure
that it is in the proper context for new uses. Hundreds of prebuilt transformation
functions combine, restructure, and aggregate information.

Transformation functionality is broad and flexible, to meet the requirements of

varied integration scenarios. For example, InfoSphere Information Server
provides inline validation and transformation of complex data types, such as the
US Health Insurance Portability and Accountability Act (HIPAA), and high-speed
joins and sorts of heterogeneous data. InfoSphere Information Server also
provides high-volume, complex data transformation and movement functionality
that can be used for stand-alone extract-transform-load (ETL) scenarios, or as a
real-time data-processing engine for applications or processes.

Deliver the information

InfoSphere Information Server virtualize, synchronize, or move information to the
people, processes, or applications that need it. Information can be delivered by
using federation or time-based or event-based processing, moved in large bulk
volumes from location to location, or accessed in place when it cannot be
consolidated.

InfoSphere Information Server provides direct, native access to various

mainframe and distributed information sources. It provides access to databases,
files, services, and packaged applications, and to content repositories and
collaboration systems. Companion products allow high-speed replication,
synchronization and distribution across databases, change data capture, and
event-based publishing of information.

Chapter 3. Overview of IBM InfoSphere DataStage 49

3.2 IBM InfoSphere Information Server Architecture
IBM Information Server combines the technologies of key information integration
functions within the IBM Information Platform & Solutions portfolio into a single
unified platform that enables companies to understand, cleanse, transform, and
deliver trustworthy and context-rich information, as shown in Figure 3-1 on
page 48.

IBM InfoSphere Information Server provides a unified architecture that works

with all types of information integration. Common services, unified parallel
processing, and unified metadata are at the core of the server architecture.

The architecture is service-oriented, enabling IBM InfoSphere Information Server

to work within evolving enterprise service-oriented architectures (SOAs). An
SOA also connects the individual suite product modules of InfoSphere
Information Server.

By eliminating the duplication of functions, the architecture efficiently uses

hardware resources and reduces the amount of development and administrative
effort that is required to deploy an integration solution.

IBM Information Server is a client-server architecture made up of client-based

design, administration, and operation tools that access a set of server-based
data integration capabilities through a common services layer, as shown in
Figure 3-2 on page 51. Figure 3-2 on page 51 is a slightly different and more
detailed view of the same information that was shown in Figure 3-1 on page 48.
Additionally, in Figure 3-2 on page 51, notice the Repository and Engine tiers. In
the following sections, we discuss these tiers.

50 InfoSphere DataStage for Enterprise XML Data Integration

 Administrative Cl ients
Platform Services
Desktop & Web Common Product-Specific
Services Services Metadata
User Clients Repository

Application Server
Desktop & Web

DataStage/QualityStage Scratch

Information Analyzer IADB

Administrative Cl ients Information Server Engine QualityStage Match DB
Desktop & Web
Connectors, Packs, Service
QualityStage Modules Agents
Communication Agents

Figure 3-2 Information Server technical components

Unified parallel processing engine

Much of the work that InfoSphere Information Server performs takes place within
the parallel processing engine. The engine handles data processing needs as
diverse as performing the analysis of large databases for IBM InfoSphere
Information Analyzer, data cleansing for IBM InfoSphere QualityStage, and
complex transformations for IBM InfoSphere DataStage. This parallel processing
engine is designed to deliver the following benefits:
򐂰 Parallelism and data pipelining to complete increasing volumes of work in
decreasing time windows
򐂰 Scalability by adding hardware (for example, processors or nodes in a grid)
with no changes to the data integration design
򐂰 Optimized database, file, and queue processing to handle large files that
cannot fit in memory all at once or large numbers of small files

Common connectivity
InfoSphere Information Server connects to information sources whether they are
structured, unstructured, on the mainframe, or applications. Metadata-driven

Chapter 3. Overview of IBM InfoSphere DataStage 51

 connectivity is shared across the suite components, and connection objects are
reusable across functions.

Connectors provide the design-time importing of metadata, data browsing and

sampling, runtime dynamic metadata access, error handling, and high
functionality and high performance runtime data access. Prebuilt interfaces for
packaged applications called packs provide adapters to SAP, Siebel, Oracle, and
other software, enabling integration with enterprise applications and associated
reporting and analytical systems.

Unified metadata
InfoSphere Information Server is built on a unified metadata infrastructure that
enables shared understanding between business and technical domains. This
infrastructure reduces development time and provides a persistent record that
can improve confidence in information. All functions of InfoSphere Information
Server share the metamodel, making it easier for different roles and functions to
collaborate.

A common metadata repository provides persistent storage for all InfoSphere

Information Server suite components. All of the products depend on the
repository to navigate, query, and update metadata. The repository contains two
kinds of metadata:
򐂰 Dynamic metadata includes design-time information.
򐂰 Operational metadata includes performance monitoring, audit and log data,
and data-profiling sample data.

Because the repository is shared by all suite components, profiling information

that is created by InfoSphere Information Analyzer is instantly available to users
of InfoSphere DataStage and InfoSphere QualityStage.

The repository is a Java 2 Platform, Enterprise Edition (J2EE) application that

uses a standard relational database, such as IBM DB2, Oracle, or SQL Server
for persistence. (DB2 is provided with InfoSphere Information Server.) These
databases provide backup, administration, scalability, parallel access,
transactions, and concurrent access.

Common services
InfoSphere Information Server is built entirely on a set of shared services that
centralizes core tasks across the platform. These tasks include administrative
tasks, such as security, user administration, logging, and reporting. Shared
services allow these tasks to be managed and controlled in one place,
regardless of which suite component is used. The common services also include
the metadata services, which provide standard service-oriented access and
analysis of metadata across the platform. In addition, the common Services tier

52 InfoSphere DataStage for Enterprise XML Data Integration

 manages how services are deployed from any of the product functions. The
common Services tier allows the cleansing and transformation rules or federated
queries to be published as shared services within an SOA with a consistent and
easy-to-use mechanism.

InfoSphere Information Server products can access three general categories of

service:
򐂰 Design services help developers create function-specific services that can
also be shared. For example, InfoSphere Information Analyzer calls a column
analyzer service that was created for enterprise data analysis. This column
analyzer service can be integrated with other parts of InfoSphere Information
Server, because it exhibits common SOA characteristics.
򐂰 Execution services include logging, scheduling, monitoring, reporting,
security, and web framework.
򐂰 Metadata services enable metadata to be shared across tools so that
changes made in one InfoSphere Information Server component are instantly
visible across all of the suite components. Metadata services are integrated
with the metadata repository. Metadata services also allow the exchange of
metadata with external tools.

The common Services tier is deployed on J2EE-compliant application servers,

such as IBM WebSphere Application Server, which is included with InfoSphere
Information Server.

Unified user interface

The face of InfoSphere Information Server is a common graphical interface and
tool framework. Shared interfaces, such as the IBM InfoSphere Information
Server Console and the IBM InfoSphere Information Server Web Console,
provide a common interface, visual controls, and user experience across
products. Common functions, such as catalog browsing, metadata import, query,
and data browsing, all expose underlying common services in a uniform way.
InfoSphere Information Server provides rich client interfaces for highly detailed
development work and thin clients that run in web browsers for administration.

Application programming interfaces (APIs) support various interface styles that

include standard request-reply, service-oriented, event-driven, and scheduled
task invocation.

3.3 Information Server component overview

IBM Information Server is a client-server architecture made up of client-based
design, administration, and operation tools that access a set of server-based

Chapter 3. Overview of IBM InfoSphere DataStage 53

 data integration capabilities through a common services layer, as shown in
Figure 3-3. Additionally, in Figure 3-3, notice the four tiers of IBM Information
Server. The Client tier provides a number of client interfaces, optimized to
different user roles within an organization. The Client tier includes IBM
InfoSphere DataStage and QualityStage clients (Administrator, Designer, and
Director), IBM Information Server Console, and IBM Information Server Web
Console.

This tier can be categorized in two broad categories of clients: Administrative

clients and User clients. Information Server provides a common layer of services
through the Services tier where all the metadata is stored in a common
repository in the Metadata Repository tier. The Engine tier provides the
high-performance parallel engine that performs analysis, cleansing, and
transformation processing. In the following sections, we explain these tiers.

Client Tier

Services Tier Engine Tier

Services for IBM IBM Information Server

Information Server Engine

Common Services
Connectors Packs

InfoSphere QualityStage
Product-specific services
modules

Service Agents
Application Ser ver (Log ging, Co mmun ications
(ASB), Job Mo nitor, Resource
T racker )

Metadata Repository Tier

Figure 3-3 Information Server tiers

54 InfoSphere DataStage for Enterprise XML Data Integration

3.3.1 Administrative clients
With these clients, you can manage the areas of security, licensing, logging, and
scheduling.

Web Console
Administration tasks are performed in the IBM Information Server Web Console.
The IBM Information Server Web Console is a browser-based interface for
administrative activities, such as managing security and creating views of
scheduled tasks.

IBM InfoSphere DataStage Administrator

For IBM InfoSphere DataStage and IBM InfoSphere QualityStage project
administration, you use the IBM InfoSphere DataStage Administrator client. It
administers IBM InfoSphere DataStage and QualityStage projects and conducts
daily operations on the server. It is used to specify general server defaults, add
and delete projects, and to set project properties. User and group privileges are
also set with the Administrator client.

3.3.2 User clients

These clients help perform client tasks, such as creating, managing, and
designing jobs, as well as profiling and analyzing data. Information Server
includes the following product modules.

IBM InfoSphere DataStage

IBM InfoSphere DataStage enables organizations to design data flows that
extract information from multiple source systems, transform it in ways that make it
more valuable, and then deliver it to one or more target databases or
applications.

IBM InfoSphere QualityStage

Designed to help organizations understand and improve the overall quality of
their data assets, IBM InfoSphere QualityStage provides advanced features to
help investigate, repair, consolidate, and validate heterogeneous data within an
integration workflow.

IBM InfoSphere Information Services Director

IBM Information Server provides a unified mechanism for publishing and
managing shared SOA services across data quality, data transformation, and
federation functions. This mechanism allows information specialists to easily
deploy services for any information integration task and consistently manage

Chapter 3. Overview of IBM InfoSphere DataStage 55

 them. Developers can take data integration logic that was built with IBM
Information Server and publish it as an “always on” service in minutes. The
common services also include the metadata services, which provide standard
service-oriented access and analysis of metadata across the platform.

IBM InfoSphere Information Analyzer

IBM InfoSphere Information Analyzer profiles and analyzes data so that you can
deliver trusted information to your users. It can automatically scan samples of
your data to determine their quality and structure. This analysis aids you in
understanding the inputs to your integration process, ranging from individual
fields to high-level data entities. With information analysis, you can correct
problems with structure or validity before it affects your project. While analysis of
source data is a critical first step in any integration project, you must continually
monitor the quality of the data. With IBM InfoSphere Information Analyzer, you
can treat profiling and analysis as an ongoing process and create business
metrics that you can run and track over time.

IBM Information Server FastTrack

IBM Information Server FastTrack simplifies and streamlines communication
between the business analyst and developer by capturing business requirements
and automatically translating them into IBM InfoSphere DataStage ETL jobs.

IBM InfoSphere Business Glossary

IBM Information Server provides a web-based tool that enables business
analysts and subject-matter experts to create, manage, and share a common
enterprise vocabulary and classification system. IBM InfoSphere Business
Glossary enables users to link business terms to more technical artifacts that are
managed by the metadata repository. The metadata repository also enables
sharing of the business terms by IBM InfoSphere Data Architect and IBM
InfoSphere Information Analyzer, creating a common set of semantic tags for
reuse by data modelers, data analysts, business analysts, and users.

IBM InfoSphere Discovery

IBM InfoSphere Discovery analyzes the data values and patterns from one or
more sources, to capture these hidden correlations and bring them clearly into
view. InfoSphere Discovery applies heuristics and sophisticated algorithms to
perform a full range of data analysis techniques: single-source and cross-source
data overlap and relationship analysis, advanced matching key discovery, and
transformation logic discovery. InfoSphere Discovery accurately identifies
relationships and business objects, accelerating the deployment of
information-centric projects.

56 InfoSphere DataStage for Enterprise XML Data Integration

 IBM InfoSphere Data Architect
IBM InfoSphere Data Architect is a data modeling tool to create and manage
your enterprise data models. Data Architect comes with a full set of features with
which you perform mappings, relationships between elements, compare and
synchronize different models, enforce standards for naming, meaning, values,
and relationships. Data Architect is a tool that is fully integrated with a suite of
applications that covers the entire software development lifecycle.

IBM Metadata Workbench

With IBM Metadata Workbench (MWB), you can explore and analyze technical,
business, and operational metadata about sources of data. IBM Metadata
Workbench helps you to graphically understand these relationships from the
sources of information to the places where information is used, even across
different tools and technologies. IBM Metadata Workbench outlines the complete
lineage of fields from applications, reports, or data warehouses back to source
systems, including the types of processing that were performed.

IBM InfoSphere Blueprint Director

With IBM InfoSphere Blueprint Director, you can define and manage a blueprint
of your information project landscape from initial sketches through delivery. The
integration team can collaborate on actionable information blueprints that
connect the business vision with the corresponding technical metadata. It directs
the team to apply preferred practices that are based on reference architectures
and methodology, provides consistency and connectivity to your information
architecture solution by linking the solution overview and detailed design
documents together, allowing all team members to understand the project as it
evolves. InfoSphere Blueprint Director includes several major capabilities that
enable governance teams to develop a common information blueprint, creating
project blueprints so that you can link blueprint elements to related artifacts.

3.3.3 Server tiers

IBM Information Server is built entirely on a set of shared services that
centralizes core tasks across the platform. Shared services allow these tasks to
be managed and controlled in one place, regardless of which suite component is
used. The Server tiers of the Information Server Platform include the Services,
Engine, and Repository tiers.

Chapter 3. Overview of IBM InfoSphere DataStage 57

 Services tier
The Services tier includes both common and product-specific services:
򐂰 Common services are used across the Information Server suite for tasks,
such as security, user administration, logging, reporting, metadata, and
execution.
򐂰 Product-specific services provide tasks for specific products within the
Information Server suite. For example, IBM InfoSphere Information Analyzer
calls a column analyzer service (a product-specific service) that was created
for enterprise data analysis. The shared service environment allows
integration across IBM Information Server, because the shared services are
deployed by using common SOA standards.

IBM Information Server products can access three general categories of service:
򐂰 Design: Design services help developers create function-specific services
that can also be shared.
򐂰 Execution: Execution services include logging, scheduling, monitoring,
reporting, security, and web framework.
򐂰 Metadata: Using metadata services, metadata is shared “live” across tools so
that changes made in one IBM Information Server component are instantly
visible across all of the suite components. Metadata services are tightly
integrated with the common repository. You can also exchange metadata with
external tools by using metadata services.

The common services layer is deployed on the J2EE-compliant application

server IBM WebSphere Application Server, which is included with IBM
Information Server.

Repository tier
The shared repository is used to store all IBM Information Server product module
objects1 (including IBM InfoSphere DataStage objects) and is shared with other
applications in the suite. Clients can access metadata and the results of data
analysis from the service layers.

Engine tier
This tier is the parallel runtime engine that executes the IBM Information Server
tasks. It consists of the Information Server Engine, Service Agents, Connectors,
and Packaged Application Connectivity Kits (PACKs2):
1
IBM Information Server product module objects include jobs and table definitions, as well as
operational metadata, such as job start and stop times. The repository is also used to store
Information Server configuration settings, such as user group assignments and roles.
2
PACKs provide an application-specific view of data and use the packaged application vendor’s
application programming interfaces (APIs) for connectivity and business metadata.

58 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 The IBM Information Server Engine consists of the products that you install,
such as IBM InfoSphere DataStage and IBM InfoSphere QualityStage. It runs
jobs to extract, transform, load, and standardize data. The engine runs
DataStage and QualityStage jobs. It also executes the parallel jobs for
Information Analyzer tasks.
򐂰 Service Agents are Java processes that run in the background on each
computer that hosts IBM InfoSphere DataStage. They provide the
communication between the Services and Engine tiers of Information Server.
򐂰 Connectors and PACKs. IBM Information Server connects to various
information sources whether they are structured, unstructured, on the
mainframe, or applications. Metadata-driven connectivity is shared across the
suite components, and connection objects are reusable across functions.
Connectors provide design-time importing of metadata, data browsing and
sampling, runtime dynamic metadata access, error handling, and high
functionality and high performance runtime data access. Prebuilt interfaces
for packaged applications called PACKs provide adapters to SAP, Siebel,
Oracle, and other software, enabling integration with enterprise applications
and associated reporting and analytical systems.
򐂰 Working areas: These working areas are temporary storage areas used by
the suite components.
򐂰 Information Services Director (ISD) Resource Providers: Information service
providers are the (data) sources of operations for your services. Using IBM
InfoSphere Information Services Director, you can create services from five
sources:
– IBM InfoSphere DataStage and QualityStage
– IBM DB2 for Linux, UNIX and Windows
– IBM InfoSphere Federation Server
– IBM InfoSphere Classic Federation Server for z/OS
– Oracle Database Server

3.3.4 Supported topologies

IBM Information Server is built on a highly scalable parallel software architecture
that delivers high levels of throughput and performance. For maximum scalability,
integration software must do more than run on Symmetric Multiprocessing (SMP)
and Massively Parallel Processing (MPP) computer systems. If the data
integration platform does not saturate all of the nodes of the MPP box or system
in the cluster or grid, scalability cannot be maximized. The IBM Information
Server components fully exploit SMP, clustered, grid, and MPP environments to
optimize the use of all available hardware resources.

Chapter 3. Overview of IBM InfoSphere DataStage 59

 IBM Information Server supports multiple topologies to satisfy various data
integration and hardware business requirements:
򐂰 Two tier
򐂰 Three tier
򐂰 Cluster
򐂰 Grid

For all topologies, you can add clients and engines (for scalability) on additional
computers.

To select a topology, you must consider your performance needs by reviewing the
capacity requirements for the topology elements: the server, disk, network, data
sources, targets, data volumes, processing requirements, and any service-level
agreements (SLAs).

We describe each topology briefly.

Tip: We suggest that you use the same topology for your test and production
environments to minimize issues when a job is deployed into production.

Microsoft Windows platform: On a Microsoft Windows platform, the clients,

engine, application server, and metadata repository can be collocated on the
same machine. This topology (not shown here) is only suitable for
demonstrations, as an educational or proof-of-concept (POC) platform.

Two tier
The engine, application server, and metadata repository are all on the same
computer system, while the clients are on a different machine, as shown in
Figure 3-4 on page 61.

High availability and failover are simpler to manage with two computers, because
all the servers fail over at the same time.

60 InfoSphere DataStage for Enterprise XML Data Integration

 Client Documentation

WebSphere
Application Server

Domain

DB2
Engine
Metadata
Repository

Figure 3-4 Two-tier topology

Three tier
The engine is on one machine, and the application server and metadata
repository are co-located on another machine. The clients are on a third
machine, as shown in Figure 3-5 on page 62.

Failover configuration is more complex, because of the increased number of

failover scenarios that are required by three or more computers.

Chapter 3. Overview of IBM InfoSphere DataStage 61

 Client Documentation

WebSphere
Application Server

Domain

DB2
Engine
Metadata
Repository

Figure 3-5 Three-tier topology

Cluster
This cluster topology is a slight variation of the three-tier topology with the engine
duplicated over multiple computers, as shown in Figure 3-6 on page 63.

In a cluster environment, a single parallel job execution can span multiple

computers, each with its own engine.

The processing of a job on the multiple machines is driven by a configuration file

that is associated with the job. The configuration file specifies the machines to be
used by the job.

62 InfoSphere DataStage for Enterprise XML Data Integration

 Client Documentation

WebSphere
Application Server
Engine

Domain

DB2

Metadata Engine
Repository

Figure 3-6 Cluster and grid

Grid
With hardware computing power a commodity, grid computing provides more
processing power to a task than was previously possible. Grid computing uses all
of the low-cost computing resources, processors, and memory that are available
on the network to create a single system image.

Grid topology is similar to that of a cluster (Figure 3-6) with engines distributed
over multiple machines. As in the case of a cluster environment, a single parallel
job execution can span multiple computers, each with its own engine.

The key difference with cluster computing is that in a grid environment, the
machines over which a job executes are dynamically determined (through the
generation of a dynamic configuration file) with an integrated resource manager,
such as IBM Tivoli® Workload Scheduler LoadLeveler®.

The parallel processing architecture of IBM Information Server uses the

computing power of grid environments and greatly simplifies the development of
scalable integration systems that run in parallel for grid environments.

Chapter 3. Overview of IBM InfoSphere DataStage 63

3.4 InfoSphere Information Server integration scenarios
Information integration is a complex activity that affects every part of an
organization. To address the most common integration business problems, these
integration scenarios show how you can deploy and use IBM InfoSphere
Information Server and the InfoSphere Foundation Tools components together in
an integrated fashion. The integration scenarios focus on data quality within a
data warehouse implementation.

3.4.1 Data integration challenges

Today, organizations face a wide range of information-related challenges:
򐂰 Varied and often unknown data quality problems
򐂰 Disputes over the meaning and context of information
򐂰 Management of multiple complex transformations
򐂰 Use of existing integration processes rather than a duplication of effort
򐂰 Ever-increasing quantities of data
򐂰 Shrinking processing windows
򐂰 Growing need for monitoring and security to ensure compliance with national
and international law

Organizations must streamline and connect information and systems across

enterprise domains with an integrated information infrastructure. Disconnected
information leaves IT organizations unable to respond rapidly to new information
requests from business users and executives. With few tools or resources to
track the information sprawl, it is also difficult for businesses to monitor data
quality and consistently apply business rules. As a result, information remains
scattered across the enterprise under a multitude of disorganized categories and
incompatible descriptions.

The following examples show several key data integration issues:

򐂰 Enterprise application source metadata is not easily assembled in one place
to understand what is available. The mix can also include existing sources,
which often do not make metadata available through a standard API, if at all.
򐂰 Master reference data, names and addresses of suppliers and customers,
and part numbers and descriptions differ across applications and duplicate
sources of this data.
򐂰 Hundreds of extract, transform, and load (ETL) jobs need to be written to
move data from all the sources to the new target application.
򐂰 Data transformations are required before loading the data so that it fits into
the new environment structures.

64 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 The ability to handle large amounts of data that can be run through the
process, and finish on time, is essential. Companies need the infrastructure to
support running any of the transformation and data-matching routines on
demand.

3.4.2 Building a data integration application scenario

Information integration is the process of integrating and transforming data and
content to deliver authoritative, consistent, timely and complete information, and
governing its quality throughout its lifecycle. The InfoSphere Information Server
platform and InfoSphere DataStage (DS) are core to these activities.

IBM InfoSphere Information Server is a software platform that helps

organizations derive more value from the complex, heterogeneous information
that is spread across their systems. It provides breakthrough collaboration,
productivity, and performance for cleansing, transforming, and moving this
information consistently and securely throughout the enterprise. It can then be
accessed and used in new ways to drive innovation, increase operational
efficiency, and lower risk.

IBM InfoSphere DataStage integrates data across multiple and high volume data
sources and target applications. It integrates data on demand with a high
performance parallel framework, extended metadata management, and
enterprise connectivity. DataStage supports the collection, integration, and
transformation of large volumes of data, with data structures that range from
simple to highly complex.

DataStage can manage data that arrives in real time, as well as data received on
a periodic or scheduled basis. This capability enables companies to solve
large-scale business problems through high-performance processing of massive
data volumes. By using the parallel processing capabilities of multiprocessor
hardware platforms, IBM InfoSphere DataStage Enterprise Edition can scale to
satisfy the demands of ever-growing data volumes, stringent real-time
requirements, and ever-shrinking batch windows.

Along with these key components, establishing consistent development

standards helps to improve developer productivity and reduce ongoing
maintenance costs. Development standards can also make it easier to integrate
external processes, such as automated auditing and reporting, and to build
technical and support documentation.

With these components and a great set of standard practices, you are on your
way to a highly successful data integration effort.

Chapter 3. Overview of IBM InfoSphere DataStage 65

 IBM InfoSphere Information Server features a unified suite of product modules
designed to streamline the process of building a data integration application.
With InfoSphere Information Server, you can perform five key integration
functions:
򐂰 Understand the data. InfoSphere Information Server helps you to
automatically discover, model, define, and govern information content and
structure, as well as understand and analyze the meaning, relationships, and
lineage of information. With these capabilities, you can better understand data
sources and relationships and define the business rules that help eliminate
the risk of using or proliferating bad data.
򐂰 Cleanse the data. InfoSphere Information Server supports information quality
and consistency by standardizing, validating, matching, and merging data.
The platform can help you create a single, comprehensive, accurate view of
information by matching records across or within data sources.
򐂰 Transform data into information. InfoSphere Information Server transforms
and enriches information to help ensure that it is in the proper context for new
uses. It also provides high-volume, complex data transformation and
movement functionality that can be used for stand-alone ETL scenarios or as
a real-time data processing engine for applications or processes.
򐂰 Deliver the right information at the right time. InfoSphere Information Server
can virtualize, synchronize, or move information to the people, processes, or
applications that need it. It also supports critical SOAs by allowing
transformation rules to be deployed and reused as services across multiple
enterprise applications.
򐂰 Perform unified metadata management. InfoSphere Information Server is built
on a unified metadata infrastructure that enables shared understanding
between the user roles involved in a data integration project, including
business, operational, and technical domains. This common, managed
infrastructure helps reduce development time and provides a persistent
record that can improve confidence in information while helping to eliminate
manual coordination efforts.

3.5 Introduction to InfoSphere DataStage

In its simplest form, IBM InfoSphere DataStage performs data transformation and
movement from source systems to target systems in batch and in real time. The
data sources might include indexed files, sequential files, relational databases,
archives, external data sources, enterprise applications, and message queues.

DataStage manages data that arrives and data that is received on a periodic or
scheduled basis. It enables companies to solve large-scale business problems

66 InfoSphere DataStage for Enterprise XML Data Integration

 with high-performance processing of massive data volumes. By using the parallel
processing capabilities of multiprocessor hardware platforms, DataStage can
scale to satisfy the demands of ever-growing data volumes, stringent real-time
requirements, and ever-shrinking batch windows.

Using the combined suite of IBM Information Server offerings, DataStage can
simplify the development of authoritative master data by showing where and how
information is stored across source systems. DataStage can also consolidate
disparate data into a single, reliable record, cleanse and standardize information,
remove duplicates, and link records across systems. This master record can be
loaded into operational data stores, data warehouses, or master data
applications, such as IBM Master Data Management (MDM), by using IBM
InfoSphere DataStage.

IBM InfoSphere DataStage delivers four core capabilities:

򐂰 Connectivity to a wide range of mainframe, existing, and enterprise
applications, databases, file formats, and external information sources.
򐂰 Prebuilt library of more than 300 functions, including data validation rules and
complex transformations.
򐂰 Maximum throughput by using a parallel, high-performance processing
architecture.
򐂰 Enterprise-class capabilities for development, deployment, maintenance, and
high availability. It uses metadata for analysis and maintenance. It also
operates in batch, real time, or as a Web service.

IBM InfoSphere DataStage enables part of the information integration process,

data transformation, as shown in Figure 3-2 on page 51.

In the following sections, we briefly describe the following aspects of IBM

InfoSphere DataStage:
򐂰 Data transformation
򐂰 Jobs
򐂰 Parallel processing

3.5.1 Data transformation

Data transformation and movement are the processes by which source data is
selected, converted, and mapped to the format that is required by targeted
systems. The processes manipulate data to bring it into compliance with
business, domain, and integrity rules and with other data in the target
environment.

Chapter 3. Overview of IBM InfoSphere DataStage 67

 Transformation can take several of the following forms:
򐂰 Aggregation: Consolidating or summarizing data values into a single value.
Collecting daily sales data to be aggregated to the weekly level is a common
example of aggregation.
򐂰 Basic conversion: Ensuring that data types are correctly converted and
mapped from source to target columns.
򐂰 Cleansing: Resolving inconsistencies and fixing the anomalies in source data.
򐂰 Derivation: Transforming data from multiple sources by using a complex
business rule or algorithm.
򐂰 Enrichment: Combining data from internal or external sources to provide
additional meaning to the data.
򐂰 Normalizing: Reducing the amount of redundant and potentially duplicated
data.
򐂰 Combining: The process of combining data from multiple sources via parallel
Lookup, Join, or Merge operations.
򐂰 Pivoting: Converting records in an input stream to many records in the
appropriate table in the data warehouse or data mart.
򐂰 Sorting: Grouping related records and sequencing data based on data or
string values.

3.5.2 Jobs
An IBM InfoSphere DataStage job consists of individual, linked stages, which
describe the flow of data from a data source to a data target.

A stage usually has at least one data input and one data output. However, certain
stages can accept more than one data input, and output to more than one stage.
Each stage has a set of predefined and editable properties that tell it how to
perform or process data. Properties might include the file name for the
Sequential File stage, the columns to sort, the transformations to perform, and
the database table name for the DB2 stage. These properties are viewed or
edited by using stage editors. Stages are added to a job and linked by using the
Designer. Figure 3-7 on page 69 shows several of the stages and their iconic
representations.

68 InfoSphere DataStage for Enterprise XML Data Integration

 Icon Stage Description
Transformer stage Performs any required conversions on
an input data set, and then passes the
data to another processing stage or to a
stage that writes data to a target
database or file.
Sort stage Performs complex high-speed sort
operations.

Aggregator stage Classifies data rows from a single input

data set into groups and computes totals
or aggregations for each group.

Complex Flat File stage Extracts data from a flat file containing
complex data structures, such as arrays
or groups.

DB2 stage Reads data from or writes data to IBM

DB2.

Figure 3-7 Stage examples

Stages and links can be grouped in a shared container. Instances of the shared
container can then be reused in different parallel jobs. You can also define a local
container within a job, which groups stages and links into a single unit. The local
container must be used within the job in which it is defined.

Separate types of jobs have separate stage types. The stages that are available
in the Designer depend on the type of job that is currently open in the Designer.

Parallel Job stages are organized into groups on the Designer palette:
򐂰 General includes stages, such as Container and Link.
򐂰 Data Quality includes stages, such as Investigate, Standardize, Reference
Match, and Survive.

Data Quality: This group applies when IBM InfoSphere QualityStage is

installed.

򐂰 Database includes stages, such as Classic Federation, DB2 UDB, DB2

UDB/Enterprise, Oracle, Sybase, SQL Server, Teradata, Distributed
Transaction, and Open Database Connectivity (ODBC).

Chapter 3. Overview of IBM InfoSphere DataStage 69

 򐂰 Development/Debug includes stages, such as Peek, Sample, Head, Tail, and
Row Generator.
򐂰 File includes stages, such as Complex Flat File, Data Set, Lookup File Set,
and Sequential File.
򐂰 Processing includes stages, such as Aggregator, Copy, FTP, Funnel, Join,
Lookup, Merge, Remove Duplicates, Slowly Changing Dimension, Surrogate
Key Generator, Sort, and Transformer.
򐂰 Real Time includes stages, such as Web Services Transformer, WebSphere
MQ, and Web Services Client.
򐂰 Restructure includes stages, such as Column Export and Column Import.

Stages: For details about all the available stages, refer to IBM InfoSphere
DataStage and QualityStage Parallel Job Developer’s Guide, SC18-9891-03,
and relevant connectivity guides for the stages that relate to connecting to
external data sources and data targets.

3.5.3 Parallel processing

Figure 3-8 represents a simple job: a data source, a Transformer (conversion)
stage, and the data target. The links between the stages represent the flow of
data into or out of a stage.

In a parallel job, each stage normally (but not always) corresponds to a process.
You can have multiple instances of each process to run on the available
processors in your system.

Data Source Transformer Data Target

Figure 3-8 Simple IBM InfoSphere DataStage job

A parallel DataStage job incorporates two types of parallel processing: pipeline

and partitioning. Both of these methods are used at run time by the Information
Server Engine to execute the simple job that is shown in Figure 3-8.

To the DataStage developer, this job appears the same way on your Designer
canvas, but you can optimize it through advanced properties.

Pipeline parallelism
In the Figure 3-8 example, all stages run concurrently, even in a single-node
configuration. As data is read from the Oracle source, it is passed to the

70 InfoSphere DataStage for Enterprise XML Data Integration

Transformer stage for transformation, where it is then passed to the DB2 target.
Instead of waiting for all source data to be read, as soon as the source data
stream starts to produce rows, these rows are passed to the subsequent stages.
This method is called pipeline parallelism. All three stages in our example
operate simultaneously regardless of the degree of parallelism of the
configuration file. The Information Server Engine always executes jobs with
pipeline parallelism.

If you ran the example job on a system with multiple processors, the stage
reading starts on one processor and starts filling a pipeline with the data that it
read. The Transformer stage starts running as soon as data exists in the pipeline,
processes it, and starts filling another pipeline. The stage that writes the
transformed data to the target database similarly starts writing as soon as data is
available. Thus, all three stages operate simultaneously.

Important: You do not need multiple processors to run in parallel. A single

processor can run multiple concurrent processes.

Partition parallelism
When large volumes of data are involved, you can use the power of parallel
processing to your best advantage by partitioning the data into a number of
separate sets, with each partition handled by a separate instance of the job
stages. Partition parallelism is accomplished at run time, instead of a manual
process that is required by traditional systems.

The DataStage developer needs to specify only the algorithm to partition the
data, not the degree of parallelism or where the job executes. By using partition
parallelism, the same job effectively is run simultaneously by several processors,
and each processor handles a separate subset of the total data. At the end of the
job, the data partitions can be collected again and written to a single data source,
as shown in Figure 3-9.

Figure 3-9 Conceptual representation of a job that uses partition parallelism

Combining pipeline and partition parallelism

The Information Server Engine combines pipeline and partition parallel
processing to achieve even greater performance gains. In this scenario, stages

Chapter 3. Overview of IBM InfoSphere DataStage 71

 process partitioned data and fill pipelines so that the next stage can start on that
partition before the previous stage finishes, as shown in Figure 3-10.

Figure 3-10 Pipeline and partition parallelism

In certain circumstances, you might want to repartition your data between stages.
This repartition might happen, for example, where you want to group data
differently. Suppose that you initially processed data based on customer last
name, but now you want to process data grouped by zip code. You must
repartition to ensure that all customers that share a zip code are in the same
group. With DataStage, you can repartition between stages as and when
necessary. With the Information Server Engine, repartitioning happens in
memory between stages, instead of writing to disk.

3.6 DataStage architecture

IBM InfoSphere DataStage facilitates data integration in both high-volume batch
and services-oriented deployment scenarios required by enterprise system
architectures. As part of the integrated IBM Information Server platform, it is
supported by a broad range of shared services and benefits from the reuse of
several suite components.

IBM InfoSphere DataStage and IBM InfoSphere QualityStage share the

infrastructure for importing and exporting data, designing, deploying, and running
jobs, and reporting. The developer uses the same design canvas to specify the
flow of data from preparation to transformation and delivery.

Multiple discrete services give IBM InfoSphere DataStage the flexibility to match
increasingly varied customer environments and tiered architectures. Figure 3-2
on page 51 shows how IBM InfoSphere DataStage Designer (labeled “User

72 InfoSphere DataStage for Enterprise XML Data Integration

 Clients”) interacts with other elements of the IBM Information Server platform to
deliver enterprise data analysis services.

In this section, we briefly describe the following topics:

򐂰 Shared components
򐂰 Runtime architecture

3.6.1 Shared components

On Figure 3-2 on page 51, the following suite components are shared between
IBM InfoSphere DataStage and IBM Information Server. For the unified user
interface, the following client applications make up the IBM InfoSphere
DataStage user interface:
򐂰 IBM InfoSphere DataStage and QualityStage Designer: A graphical design
interface is used to create InfoSphere DataStage applications, which are
known as jobs. Because transformation is part of data quality, the InfoSphere
DataStage and QualityStage Designer is the design interface for both
InfoSphere DataStage and InfoSphere QualityStage. Each job specifies the
data sources, required transformations, and destination of the data. Jobs are
compiled to create parallel job flows and reusable components that are
scheduled by the InfoSphere DataStage and QualityStage Director and run in
parallel by the Information Server Engine. The Designer client manages
design metadata in the repository, while compiled execution data is deployed
on the Information Server Engine tier.
򐂰 IBM InfoSphere DataStage and QualityStage Director: A graphical user
interface that is used to validate, schedule, run, and monitor InfoSphere
DataStage job sequences. The Director client shows job runtime information,
including job status and detailed job logs. This client can also be used to
establish schedules for job execution.
򐂰 InfoSphere DataStage and InfoSphere QualityStage Administrator: A
graphical user interface that is used for administration tasks:
– Setting up DataStage and Information Server Engine users
– Creating, deleting, and customizing projects
– Setting up criteria for purging runtime log records

Common services
As part of the IBM Information Server Suite, DataStage uses the common
services, as well as DataStage-specific services.

Chapter 3. Overview of IBM InfoSphere DataStage 73

 The common services provide flexible, configurable interconnections among the
many parts of the architecture:
򐂰 Metadata services, such as impact analysis and search
򐂰 Execution services that support all InfoSphere DataStage functions
򐂰 Design services that support the development and maintenance of
InfoSphere DataStage tasks

Common repository
The common repository holds three types of metadata that are required to
support IBM InfoSphere DataStage:
򐂰 Project metadata: All the project-level metadata components, including IBM
InfoSphere DataStage jobs, table definitions, built-in stages, reusable
subcomponents, and routines, are organized into folders.
򐂰 Operational metadata: The repository holds metadata that describes the
operational history of integration process runs, the success or failure of jobs,
the parameters that were used, and the time and date of these events.
򐂰 Design metadata: The repository holds design-time metadata that is created
by the IBM InfoSphere DataStage and QualityStage Designer and IBM
InfoSphere Information Analyzer.

The common parallel processing engine runs executable jobs that extract,
transform, and load data in a wide variety of settings. The engine uses
parallelism and pipelining to handle high volumes of work more quickly and to
scale a single job across the boundaries of a single server in cluster or grid
topologies.

The common connectors provide connectivity to many external resources and

access to the common repository from the processing engine. Any data source
that is supported by IBM Information Server can be used as input to or output
from an IBM InfoSphere DataStage job.

3.6.2 Runtime architecture

In this section, we briefly describe the generation of the IBM Orchestrate® SHell
(OSH) script. We describe the execution flow of IBM InfoSphere DataStage using
the Information Server Engine.

OSH script
The IBM InfoSphere DataStage and QualityStage Designer client creates IBM
InfoSphere DataStage jobs that are compiled into parallel job flows, and reusable
components that execute on the parallel Information Server Engine. With the

74 InfoSphere DataStage for Enterprise XML Data Integration

InfoSphere DataStage and QualityStage Designer client, you can use familiar
graphical point-and-click techniques to develop job flows for extracting,
cleansing, transforming, integrating, and loading data into target files, target
systems, or packaged applications.

The Designer generates all the code. It generates the OSH and C++ code for any
Transformer stages used.

The Designer performs the following tasks:

򐂰 Validates link requirements, mandatory stage options, and transformer logic
򐂰 Generates OSH representation of data flows and stages (representations of
framework “operators”)
򐂰 Generates transform code for each Transformer stage, which is then compiled
into C++ and then to corresponding native operators
򐂰 Compiles Reusable BuildOp stages by using the Designer GUI or from the
command line

You need to know the following information about the OSH:

򐂰 Comment blocks introduce each operator, the order of which is determined by
the order that you added stages to the canvas.
򐂰 OSH uses the familiar syntax of the UNIX shell, such as Operator name,
schema, operator options (“-name value” format), input (indicated by n<
where n is the input#), and output (indicated by n> where n is the output #).
򐂰 For every operator, input and output datasets are numbered sequentially
starting from zero.
򐂰 Virtual datasets (in-memory, native representation of data links) are
generated to connect operators.

Execution order: The execution order of operators is dictated by input/output

designators, and not by their placement on the diagram. The datasets connect
the OSH operators. These datasets are “virtual datasets,” that is, in-memory
data flows. Link names are used in dataset names; therefore, ensure that you
use meaningful names for the links.

Framework (Information Server Engine) terms and DataStage terms have

equivalency. The GUI frequently uses terms from both paradigms. Runtime
messages use framework terminology, because the framework engine is where
execution occurs.

The following list shows the equivalency between framework and DataStage
terms:

Chapter 3. Overview of IBM InfoSphere DataStage 75

 򐂰 Schema corresponds to table definition
򐂰 Property corresponds to format
򐂰 Type corresponds to SQL type and length
򐂰 Virtual dataset corresponds to link
򐂰 Record/field corresponds to row/column
򐂰 Operator corresponds to stage
򐂰 Step, flow, and OSH command correspond to a job
򐂰 Framework corresponds to Information Server Engine

Execution flow
When you execute a job, the generated OSH and contents of the configuration
file ($APT_CONFIG_FILE) are used to compose a “score,” which is similar to an
SQL query optimization plan.

At run time, IBM InfoSphere DataStage identifies the degree of parallelism and
node assignments for each operator, and inserts sorts and partitioners as
needed to ensure correct results. It also defines the connection topology (virtual
datasets/links) between adjacent operators/stages, and inserts buffer operators
to prevent deadlocks (for example, in fork-joins). It also defines the number of
actual operating system (OS) processes. Multiple operators/stages are
combined within a single OS process as appropriate, to improve performance
and optimize resource requirements.

The job score is used to fork processes with communication interconnects for
data, message, and control. Processing begins after the job score and processes
are created. Job processing ends when either the last row of data is processed
by the final operator, a fatal error is encountered by any operator, or the job is
halted by DataStage Job Control or human intervention, such as DataStage
Director STOP.

Settings: You can direct the score to a job log by setting

$APT_DUMP_SCORE. To identify the Score dump, look for “main program:
This step....”.

You can set $APT_STARTUP_STATUS to show each step of the job start-up.
You can set $APT_PM_SHOW_PIDS to show process IDs in the DataStage
log.

Job scores are divided into two sections: datasets (partitioning and collecting)
and operators (node/operator mapping). Both sections identify sequential or
parallel processing.

The execution (orchestra) manages control and message flow across processes
and consists of the conductor node and one or more processing nodes, as

76 InfoSphere DataStage for Enterprise XML Data Integration

shown in Figure 3-11. Actual data flows from player to player. The conductor and
section leader are only used to control process execution through control and
message channels:
򐂰 Conductor is the initial framework process. It creates the Section Leader (SL)
processes (one per node), consolidates messages to the DataStage log, and
manages orderly shutdown. The Conductor node has the start-up process.
The Conductor also communicates with the players.
򐂰 Section Leader is a process that forks player processes (one per stage) and
manages up/down communications. SLs communicate between the
conductor and player processes only. For a specific parallel configuration file,
one section leader is started for each logical node.
򐂰 Players are the actual processes that are associated with the stages. A player
sends stderr and stdout to the SL, establishes connections to other players
for data flow, and cleans up on completion. Each player must be able to
communicate with every other player. Separate communication channels
(pathways) exist for control, errors, messages, and data. The data channel
does not go through the section leader/conductor, because this design limits
scalability. Data flows directly from the upstream operator to the downstream
operator.

Conductor Node

Processing Node Processing Node

Section Leader Section Leader
(one per node) (one per node)

Player Player Player Player Player Player

OS Process OS Process

Player Player

Player Player

Figure 3-11 Parallel execution flow

Chapter 3. Overview of IBM InfoSphere DataStage 77

3.6.3 XML Pack architecture
XML Pack is a DataStage connector that is designed to handle general XML data
processing. XML Pack is installed as part of DataStage installation in the Client,
Service, and Engine tiers. XML Pack architecture is depicted in Figure 3-12.

Design Environment Runtime Environment

DataStage Designer

XML Mapper
Menu Tools St age
Custom UI
Running as a Regular
Web 2.0 Clients Stage Inside
XML Metadata Assembly
DataStage Engines
Importer Editor
UI UI

Schemas and Assembly DS Engine

Service Definitions Definitions Common Connector
Information Server Domain JVM
XML Pack 3.0 App Server Jobs E 2 Connector

Restful Services

XML Tables Dat aStage jobs PX Engine

Type Cache containing Common Connector
E2 Compiler
assembly as a
JVM
string property
XMeta of the stage E 2 Connector
E2 Connect or
PX Jobs
Contracts Model Assemblies Model
(XSD Schemas + Common Model
(E2 Steps and (ASCL)
WSDL Definitions) Mappings)

Figure 3-12 XML Pack architecture

XML Pack has four logic tiers:

򐂰 Client tier has two major components:
– XML Metadata Importer UI, which is also called Schema Library Manager
– Assembly Editor, which you use to create the XML Pack job
򐂰 Service tier provides Representational State Transfer (REST) services to
wrap around the XML Pack schema type library cache and XML Pack
compilation function.

78 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 Metadata Repository tier stores schema library information and the assembly
information.
򐂰 Engine tier hosts the XML Pack connector run time to run the XML Pack job.

In terms of binary components, XML Pack provides client, services, and engine
components. The Client tier has one executable file and a set of dynamic link
library (DLL) files.

Important: Because the Client tier is installed on top of the Windows

operating system, binary components are Windows components.

The Service tier has one WAR file for client interaction and one EAR file to
provide REST service. The Engine tier has one large JAR file, which is expanded
as a list of JAR files in the engine installation directory.

3.7 Real-time applications

The parallel framework is extended for real-time applications. The parallel
framework includes three extended aspects:
򐂰 Always-on: Real-time stage types that keep jobs always up and running.
򐂰 End-of-wave: The ability for jobs to remain always-on introduces the need to
flush records across process boundaries.
򐂰 Support for transactions in target database stages.

3.7.1 Always-on source stage types

The ability for a job to stay always-on is determined by the nature of the source
stages. In batch applications, source stages read data from sources, such as flat
files, database queries, and FTP feeds. All those stages read their corresponding
sources until exhaustion. Upon reaching the end-of-file (EOF), each source
stage propagates downstream the end-of-file (EOF) to the next stage in the
pipeline. The job terminates when all target stages run to completion.

In real-time applications, this model is inefficient, because the job must remain
running even though there might not be any data for it to process. Under those
circumstances, the stages are to remain idle, waiting for data to arrive.

Chapter 3. Overview of IBM InfoSphere DataStage 79

 A few stages keep listening for more incoming data and terminate only under
certain special circumstances, depending on the type of stage:
򐂰 ISD Input
This stage is the input stage for Information Services Director applications. It
is the same source stage for a number of supported bindings, such as
Enterprise JavaBeans (EJB), SOAP, and Java Message Service (JMS).
򐂰 MQConnector:
– MQ reader stage in Information Server 8.x.
– For pre-8.x Information Server, the corresponding stage is named
MQRead.
򐂰 Custom plug-ins
New stage types can be created for special purposes with the following
approaches:
– Custom operators
– BuildOps
– Java client

The termination of real-time DS jobs is started by each source stage:

򐂰 For ISD jobs, the Information Services Director application must be
suspended or undeployed.
򐂰 For applications, the MQConnector must reach one of the following
conditions:
– It reads a message of a certain special type (configured on the stage
properties).
– It reaches a read timeout.
– It reads a certain pre-determined maximum number of messages.

3.7.2 End-of-wave
The ability for jobs to remain always-on introduces the need to flush records at
the end of each request.

An end-of-wave (EOW) marker is a hidden record type that forces the flushing of
records. An EOW marker is generated by the source real-time stage, depending
on its nature and certain conditions set as stage properties.

Figure 3-13 on page 81 shows an example of how EOW markers are propagated
through a parallel job. The source stage can be an MQConnector; in this case,
the target stage is a Distributed Transaction stage (DTS).

80 InfoSphere DataStage for Enterprise XML Data Integration

EOW markers are propagated as normal records. However, they do not carry any
data. They cause the state of stages to be reset, as well as records to be flushed
out of virtual datasets, so a response can be forced.

Figure 3-13 EOWs flowing through a parallel job

The ISD and MQConnector stages generate EOW markers in the following way:
򐂰 ISD Input:
– It issues an EOW for each incoming request (SOAP, EJB, or JMS).
– ISD converts an incoming request to one or more records. A request might
consist of an array of records, or a single record, such as one large XML
payload. The ISD input stage passes on downstream all records for a
single incoming request. After passing on the last record for a request, it
sends out an EOW marker.
– The mapping from incoming requests to records is determined when the
operation is defined by the Information Services Console.

Chapter 3. Overview of IBM InfoSphere DataStage 81

 򐂰 MQConnector:
– It issues an EOW for one or more incoming messages. The EOW is
determined by a parameter named Record Count.
– It issues an EOW after a certain amount of time elapses, which is
determined by a parameter named Time Interval.
– A combination of the two.

EOWs modify the behavior of regular stages. Upon receiving EOW markers, the
internal state of the stage must be reset, so a new execution context begins. For
most stages (Parallel Transformers, Modify, or Lookups), there is no practical
impact from a job design perspective.

For database sparse lookups, record flow for the Lookup stage is the same. But
the stage needs to keep its connections to the database across waves, instead of
reconnecting after each wave.

However, there are a few stages whose results are directly affected by EOWs:
򐂰 Sorts
򐂰 Aggregations

For these stages, the corresponding logic is restricted to the records that belong
to a certain wave. Instead of consuming all records during the entire execution of
the job, the stage produces a partial result, for the records that belong to a
certain wave. A Sort stage, for instance, writes out sorted results that are sorted
only in the wave, and not across waves. The stage continues with the records for
the next wave, until a new EOW marker arrives.

3.7.3 Transaction support

Support for real-time applications is not complete without the proper handling of
database transactions. Batch applications rely on standard database stage types
to execute bulk loads or inserts/updates/deletes.

The transaction context for database stages prior to Information Server 8.5
always involved a single target table. Those stages support a single input link,
which maps to one or a couple of SQL statements against a single target table. If
there are multiple database stages on a job, each stage works on its own
connection and transaction context. One stage is oblivious to what is going on in
other database stages, as depicted in Figure 3-14 on page 83.

82 InfoSphere DataStage for Enterprise XML Data Integration

Figure 3-14 Transaction contexts with standard database stage types

The grouping of records into transactions might be specified as a stage property.

However, from a database consistency perspective, ultimately the entire batch
process tends to be the transactional unit. If one or more jobs fail, exceptions
must be addressed and jobs must be restarted and completed so the final
database state at the end of the batch window is fully consistent.

In real-time scenarios, restarting jobs is not an option. The updates from a

real-time job must yield a consistent database after the processing of every
wave.

Prior to Version 8.1, no other option was available to Information Services

Director (ISD) and 7.X Real-Time Integration (RTI) jobs other than resorting to
multiple separate database stages when applying changes to target databases
as part of a single job flow. The best option was to synchronize the reject output
of the database stages before sending the final response to the ISD Output
stage. See 3.8.3, “Synchronizing database stages with ISD output” on page 95
for a description of this technique.

The Connector stages, introduced in Information Server 8.0.1 and further

enhanced in 8.5, provide for a new uniform interface that enables enhanced
transaction support when dealing with a single target database stage type. This
new uniform interface enables enhanced transaction support when dealing with
message-oriented and heterogeneous targets, such as the Distributed
Transaction stage (DTS).

We describe the Connector and DTS stages in the following subsections.

Chapter 3. Overview of IBM InfoSphere DataStage 83

 Connector stages
Information Server 8.5 has Connector stages for the following targets:
򐂰 DB2
򐂰 Oracle
򐂰 Teradata
򐂰 ODBC
򐂰 IBM MQSeries®

Begining with IS 8.5, Connectors support multiple input links, instead of a single
input link, which is the limit in versions prior to IS 8.5 for Connectors and
Enterprise database stage types. With multiple input links, a Connector stage
executes all SQL statements for all rows from all input links as part of single unit
of work. Figure 3-15 shows this action.

Figure 3-15 Multistatement transactional unit with a connector for a single database type

One transaction is committed per wave. It is up to a source stage (such as ISD

input or MQConnector) to generate EOW markers that are propagated down to
the target database Connector.

With database Connectors, ISD jobs no longer have to cope with potential
database inconsistencies in the event of failure. ISD requests might still need to

84 InfoSphere DataStage for Enterprise XML Data Integration

be re-executed (either SOAP, EJB, or JMS, depending on the binding type), but
the transactional consistency across multiple tables in the target database is
guaranteed as a unit of work.

For transactions spanning multiple database types and those transactions that
include guaranteed delivery of MQ messages, you must use the Distributed
Transaction stage, which is described next.

Distributed Transaction stage (DTS)

The stage that provides transactional consistency across multiple SQL
statements for heterogeneous resource manager types in message-oriented
applications is the Distributed Transaction stage (DTS).

This stage was originally introduced in DataStage 7.5, with the name
UnitOfWork. It was re-engineered for Information Server 8, and it is now called
DTS. Figure 3-16 illustrates DTS.

Figure 3-16 Distributed Transaction stage

Chapter 3. Overview of IBM InfoSphere DataStage 85

 The DTS applies to target databases all the SQL statements for incoming
records for all input links for a specific wave, as a single unit-of-work. Each wave
is independent from the other waves.

It was originally intended to work in conjunction with the MQConnector for

message-oriented processing. When used in conjunction with the MQConnector,
the DTS provides the following capabilities:
򐂰 Guaranteed delivery from a source queue to a target database
No messages are discarded without making sure that the corresponding
transactions are successfully committed.
򐂰 One time only semantics
No transaction is committed more than once.

In IS 8.5, the DTS supports transactions against single or heterogeneous

resource managers as part of the same transaction:
򐂰 DB2
򐂰 Oracle
򐂰 Teradata
򐂰 ODBC
򐂰 MQ

Also, you can use the DTS without a source MQConnector, such as in
Information Services Director jobs. You still need to install and configure MQ,
because it is required as the underlying transaction manager.

3.7.4 Transactional support in message-oriented applications

In this section, we provide an overview of transactional support in
message-oriented applications.

Information Server supports two types of solutions for message-oriented

applications:
򐂰 Information Services Director with JMS bindings
򐂰 MQ/DTS

The first solution is for interoperability with JMS. The second solution is for MQ.

You can use ISD with JMS to process MQSeries messages and MQ/DTS to
process JMS messages by setting up a bridging between MQ and JMS with
WebSphere enterprise service bus (ESB) capabilities. However, the bridging
setup is relatively complexity between JMS and MQ.

86 InfoSphere DataStage for Enterprise XML Data Integration

We put both solutions side-by-side in a diagram in Figure 3-17. The goal of this
illustration is to draw a comparison of how transactions are handled and the path
that the data flows.

Figure 3-17 Transactional contexts in ISD and MQ/DTS jobs

Both ISD and MQ/DTS jobs are parallel jobs, which are composed of processes
that implement a pipeline, possibly with multiple partitions. The parent-child
relationships between OS processes are represented by the dotted green lines.
The path that is followed by the actual data is represented by solid (red) lines.

ISD jobs deployed with JMS bindings have the active participation of the
WebSphere Application Server and the ASB agent. In an MQ/DTS job, the data
flow is restricted to the parallel framework processes.

MQ/DTS jobs provide both guaranteed delivery and once-and-only-once

semantics. A specific transaction is not committed twice against the target
database. And, each transaction is guaranteed to be delivered to the target
database. Transaction contexts are entirely managed by DS job stages. A local
transaction is managed by the MQConnector (for the transfer of messages from
the source queue to the work queue), and an XA transaction involves multiple
database targets and the work queue.

Chapter 3. Overview of IBM InfoSphere DataStage 87

 In an ISD job, transaction contexts are managed by DS parallel processes,
depending on how many database stages are present in the flow. See 3.8.3,
“Synchronizing database stages with ISD output” on page 95 through 3.8.5,
“Information Services Director with connectors” on page 98.

However, one additional transaction context exists on the JMS side, which is
managed by EJBs in the WebSphere Application Server J2EE container as Java
Transaction API (JTA) transactions.

JTA transactions make sure that no messages are lost. If any components along
the way (WebSphere Application Server, ASB Agent, or the parallel job) fail
during the processing of an incoming message before a response is placed on
the response queue, the message is reprocessed. The JTA transaction that
selected the incoming message from the source queue is rolled back, and the
message remains on the source queue. This message is then reprocessed on
job restart.

Therefore, database transactions in ISD jobs that are exposed as JMSs must be
idempotent. For the same input data, they must yield the same result on the
target database.

The MQ/DTS and ISD/JMS solutions offer these strengths:

򐂰 MQ/DTS
Guaranteed delivery from a source queue to a target database
򐂰 ISD/JMS
Adequate for request/response scenarios when JMS queues are the delivery
mechanism

The DTS is enhanced in Version 8.5 to support MQ queues as targets.

Therefore, a request/response type of scenario can be implemented with
MQ/DTS. However, this scenario implies a non-trivial effort in setting up the
necessary bridging between JMS and MQ on the WebSphere Application Server
instance.

Consider these aspects when taking advantage of ISD with JMS bindings:
򐂰 The retry attempt when JTA transactions are enabled largely depends on the
JMS provider. In WebSphere 6, with its embedded JMS support, the default is
five attempts (this number is configurable). After five attempts, the message
is considered a poison message and goes into the dead letter queue.
򐂰 Subtle differences from provider to provider can cause problems. You must
understand exactly how things operate when MQ is the provider or when the
embedded JMS in WebSphere Application Server is the provider.

88 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 The JMS binding also creates other issues because of the pool of EJB.
Multiple queue listeners can result in confusion when messages are out of
order and a flood of concurrent clients go into ISD and overwhelm the number
of instances that are established for DataStage.

3.8 Publishing jobs as services

InfoSphere Information Services Director (ISD) is used to create live, real-time
jobs. ISD automates the publication of jobs, maps, and federated queries as
services of the following types:
򐂰 SOAP
򐂰 EJB
򐂰 JMS

ISD is notable for the way that it simplifies the exposure of DS jobs as SOA
services, letting users bypass the underlying complexities of creating J2EE
services for the various binding types.

ISD controls the invocation of those services, supporting request queuing and
load balancing across multiple service providers (a DataStage Engine is one of
the supported provider types).

A single DataStage job can be deployed as different service types, and it can
retain a single data flow design. DS jobs that are exposed as ISD services are
referred to as “ISD Jobs” throughout this section.

Figure 3-18 on page 90 is reproduced from the ISD manual and depicts its major
components. The top half of the diagram shows the components that execute
inside the WebSphere Application Server on which the Information Server
Domain layer runs. Information Server and ISD are types of J2EE applications
that can be executed on top of J2EE containers.

With Version 8.5, Information Server is tightly coupled with WebSphere

Application Server, as is Information Services Director and its deployed
applications.

The bottom half of Figure 3-18 on page 90 presents components that belong to
the Engine Layer, which can reside either on the same host as the Domain, or on
separate hosts.

Each WebSphere Information Services Director (WISD) endpoint relates to one

of the possible bindings: SOAP, JMS, or EJB. Such endpoints are part of the
J2EE applications that are seamlessly installed on the Domain WebSphere

Chapter 3. Overview of IBM InfoSphere DataStage 89

 Application Server when an ISD application is successfully deployed with the
Information Server Console.

Figure 3-18 Major components of Information Services Director

The endpoints forward incoming requests to the ASB adapter, which provides
load balancing and interfacing with multiple services providers. Load balancing is
another important concept in SOA applications. In this context, it means the
spreading incoming requests across multiple DataStage engines.

A single DS engine can service a considerable number of requests. However, if

bottlenecks are identified, more engines might be added. In this case, the same
DS jobs must be deployed on all participating engines.

The ASB agent is a Java application that intermediates the communication

between the ASB adapter and the DS engine. The ASB agent is a stand-alone
Java application, because it does not run inside a J2EE container. Because it is a
Java application, it is multithreaded and supports servicing multiple incoming
requests.

An ASB agent implements the queuing and load balancing of incoming requests.
Inside a specific DS engine, multiple pre-started, always-on job instances of the

90 InfoSphere DataStage for Enterprise XML Data Integration

same type might exist, ready to service requests. The always-on nature is
important in this situation. Instead of incurring the cost of reactivating jobs
whenever a new request arrives, the ASB agent controls the lifecycle of jobs,
keeping them up and running for as long as the ISD application is supposed to be
active.

One important concept to understand is the pipelining of requests. Multiple

requests are forwarded to an ISD job instance, before responses are produced
by the job and sent back to the ASB agent. A correlation exists between this
concept and the pipeline parallelism of DS parallel jobs. Multiple concurrent
processes execute the steps of a parallel job in tandem. Pipelining requests
allows multiple requests to flow through a job at the same time.

For ISD applications, a direct mapping exists between a service request and a
wave. For each service request, an end-of-wave marker is generated. See 3.7.2,
“End-of-wave” on page 80.

In this section, we include an explanation of how the components in Figure 3-18

on page 90 map to Information Server layers and how they interact with each
other from various aspects:
򐂰 Job activation
򐂰 Parent-child process relationships
򐂰 The flow of actual data

Figure 3-19 on page 92 illustrates these relationships.

After an ISD job is compiled and ready, the ISD developer creates an operation
for that job by using the Information Server Console. That operation is created as
part of a service, which is an element of an ISD application.

When the ISD operations, services, and application are ready, you can use the
Information Server Console to deploy that application. The result is the
installation of a J2EE application on the WebSphere Application Server instance.
The deployment results in the activation of one or more job instances in the
corresponding service provider, which are the DS engines that participate in this
deployment.

This deployment is represented by the dashed (blue) arrows. The ISD

Administration application forwards the request to the target DS engines through
the ASB adapter and the ASB agents.

Chapter 3. Overview of IBM InfoSphere DataStage 91

 Figure 3-19 Data and control paths in ISD applications

The DS engine, in turn, creates one or more parallel job instances. A parallel job
is started with an OSH process (the conductor). This process parses the OSH
script that the ISD job flow was compiled into and launches the multiple section
leaders and players that implement the runtime version of the job. This process is
represented by the dashed (green) arrows.

Incoming requests follow the path that is depicted with the red arrows. They
originate from remote or local applications, such as SOAP or EJB clients, and
even messages posted onto JMS queues. One endpoint exists for each type of
binding for each operation.

All endpoints forward requests to the local ASB adapter. The ASB adapter
forwards a request to one of the participating engines according to a load
balancing algorithm. The request reaches the remote ASB agent, which puts the
request in the pipeline for the specific job instance.

The ASB agent sends the request to the WebSphere Information Services
Director (WISD) Input stage for the job instance. The ISD job instance processes
the request (an EOW marker is generated by the WISD Input for each request).
The job response is sent back to the ASB agent through the WISD Output stage.

92 InfoSphere DataStage for Enterprise XML Data Integration

 The response flows back to the caller, flowing through the same components
from which it came originally. Notice that the WebSphere Application Server
actively participates in the request processing.

3.8.1 Design topology rules for always-on ISD jobs

The following rules are the job flow design topology rules for always-on ISD jobs:
򐂰 There must be one source WISD Input stage and one target WISD Output
stage.
򐂰 The following rules must be obeyed:
– All data paths must originate from the same source stage (the WISD
Input). Reference links to Normal Lookups are the exceptions. Reference
links to Normal Lookups must originate from independent branches.
– All data paths must lead to the same target WISD Output stage.
– Dangling branches might lead to Sequential File stages, for instance, but
these dangling branches are to be used only for development and
debugging. They must not be relied on for production jobs.
– Normal Lookups can only be used for static reference tables.
– For any lookups against dynamic database tables, sparse database
lookups must be used.
– Sparse lookups might return multiple records, depending on the business
requirements.

3.8.2 Scalability
ISD supports two layers of scalability:
򐂰 Multiple job instances
򐂰 Deployment across multiple service providers, that is, multiple DS engines

These layers are illustrated in Figure 3-20 on page 94.

Chapter 3. Overview of IBM InfoSphere DataStage 93

 ISD Load
http
Balancing
App
server Server

DS Server

Job instances

DS Server
ISD Load
Balancing

ASB Agent

Figure 3-20 ISD load balancing

Whether to deploy multiple job instances and multiple DS engines to service

requests for the same ISD application is a matter of performance monitoring and
tuning. The performance of the application must be closely monitored to
understand possible bottlenecks, which then drive the decisions in terms of
additional job instances and engines.

To address performance requirements, we recommend making assessments in

the following order:
1. Job design efficiency
2. Number of job instances
3. Number of DS engines

The first task is to ensure that the logic is efficient, which includes ensuring that
database transactions, transformations, and the entire job flow are optimally
designed.

After the job design is tuned, assess the maximum number of requests that a
single job instance can handle. This maximum is a function of the job complexity
and the number of requests (in other words, EOW markers) that can flow through
the job simultaneously in a pipeline. If that number of requests is not enough to
service the amount of incoming requests, more than one job instance can be
instantiated. In most cases, increasing the number of job instances helps meet
the service-level agreement (SLA) requirements.

However, cases might occur when you reach the maximum number of requests
that a single ASB agent can handle. Therefore, the limit of a single DS engine is

94 InfoSphere DataStage for Enterprise XML Data Integration

 reached. You can verify this limit when no matter how many job instances are
instantiated, the engine cannot handle more simultaneous requests. In this
situation, add more DS engines either on the same host (if there is enough spare
capacity) or on separate hosts.

Remember that throughout this tuning exercise, we assume that you have
enough hardware resources. Increasing the number of job instances and DS
engines does not help if the CPUs, disks, and network are already saturated.

3.8.3 Synchronizing database stages with ISD output

Prior to Information Server 8.5, ISD jobs relied on multiple database stages when
updating more than one table in a single job flow, similar to what was discussed
in 3.7.3, “Transaction support” on page 82.

Figure 3-14 on page 83 illustrates database transaction contexts in a batch

application. Each separate database stage maintains a separate connection and
transaction context. One database stage is oblivious to what is going on in other
database stages even if they are part of the same job.

In an always-on job, the database stages must complete the SQL statements
before the response is returned to the caller (that is, before the result is sent to
the WISD Output stage).

Pre-8.5 jobs required a technique, which is illustrated in Figure 3-21, that

involves a sequence of stages connected to the reject link of a standard
database stage.

Figure 3-21 Synchronizing database and WISD output

The only exception to this rule in releases prior to 8.5 was the Teradata
Connector, which in Version 8.0.1 already supported an output link similar to
what is described in 3.8.5, “Information Services Director with connectors” on
page 98. However, prior to 8.5, the Teradata Connector did not allow more than a
single input link.

Chapter 3. Overview of IBM InfoSphere DataStage 95

 A Column Generator created a column, which was used as the aggregation key
in the subsequent aggregator. The output from the aggregator and the result
from the main job logic (depicted as a local container) synchronized with a Join
stage.

The Join stage guaranteed that the response was sent only after the database
statements for the wave complete (either successfully or with errors).

The logic that is implemented by the Column Generator, Aggregator, and Join
stages is repeated for each standard database stage that is present in the flow.
Synchronized results ensure that no other database activity occurs for a request
that is already answered.

Again, those steps were necessary in pre-8.5 releases. In IS 8.5, the Database
Connectors are substantially enhanced to support multiple input links and output
links that can forward not only rejected rows, but also processed rows.

Two alternatives for always-on 8.5 jobs are available for database operations:
򐂰 DTS
򐂰 Database Connector stages

We describe these alternatives in subsequent sections.

3.8.4 ISD with DTS

The DTS stage is enhanced as part of the 8.5 release to support transactions
without the presence of source and work queues. This enhancement enables a
DTS stage to respond and commit transactions solely as a result of EOW
markers. Using the mode that is described in this section, a DTS stage can be
included as part of an always-on ISD job. The DTS stage supports units of work
that meet the following requirements:
򐂰 Multiple SQL statements
򐂰 Multiple target types:
– DB2
– Oracle
– Teradata
– Oracle
– MQSeries

Figure 3-22 on page 97 illustrates a DTS stage that is used in an ISD job. It
significantly reduces the clutter that is associated with the synchronization of the
standard database stages. We provide this illustration to give you an overall
perspective.

96 InfoSphere DataStage for Enterprise XML Data Integration

The DTS stage supports an output link, whose table definition can be found in
category Table Definitions/Database/Distributed Transaction in the DS repository
tree.

When used in ISD jobs, the Use MQ Messaging DTS property must be set to NO.
Although source and work queues are not present, MQSeries must still be
installed and available locally, because it acts as the XA transaction manager.

DTS must be used in ISD jobs only when there are multiple target database
types and multiple target database instances. If all SQL statements are to be
executed on the same target database instance, a database connector must be
used instead.

Figure 3-22 DTS in an Information Services Director job

Chapter 3. Overview of IBM InfoSphere DataStage 97

3.8.5 Information Services Director with connectors
The database connectors in IS 8.5 support units of work involving multiple SQL
statements against the same target database instance.

For Connector stages, an output link carries the input link data, plus an optional
error code and error message.

If configured to output successful rows to the reject link, each output record
represents one incoming row to the stage. Output links are already supported by
the Teradata Connector in IS 8.0.1, although that connector is still restricted to a
single input link.

Figure 3-23 shows an example of multiple input links to a DB2 Connector (units
of work with Connector stage in an Information Services Director job). All SQL
statements for all input links are executed and committed as part of a single
transaction, for each wave. An EOW marker triggers the commit.

All SQL statements must be executed against the same target database
instance. If more than one target database instance is involved (of the same or
different types), the DTS stage must be used instead.

The example also depicts multiple input links to a DB2 Connector (units of work
with the Connector stage in an Information Services Director job).

Homogeneous

Figure 3-23 Units of work with the Connector stage

3.8.6 Partitioning in ISD jobs

In InfoSphere DataStage parallel jobs, when a multiple node configuration file is
used, data can be passed through different paths and at different speeds

98 InfoSphere DataStage for Enterprise XML Data Integration

between the stages of the job. Because of this architecture, the data records that
make up an Information Services Director (ISD) service request can arrive at the
ISD output stage in a different order than they were received. When this situation
occurs, the response data produced by the job for one request can be directed to
a different request. Or, the response for one request might contain both its own
data plus the data that was meant for a different request. When a multiple node
configuration file is used and multiple requests are processed simultaneously,
there is no guarantee that the response provided contains the correct data.

This issue only occurs for jobs that are deployed as ISD providers and that meet
all of the criteria in the following list:
򐂰 The job has an ISD Input stage.
򐂰 The configuration file that is used by the job (APT_CONFIG_FILE) contains
multiple nodes.
򐂰 Multiple requests can be processed by the job at the same time.

To avoid this problem, ensure that your ISD jobs only use a single-node
configuration file. You can ensure that your ISD jobs only use a single-node
configuration file by adding the environment variable $APT_CONFIG_FILE to the
job properties at the job level. You then need to set $APT_CONFIG_FILE to
reference a configuration file that contains only a single defined node. After this
change, you must disable this job within Information Services Director, recompile
it in DataStage Designer, and re-enable it in the ISD Console.

Alternatively, if all jobs in the project are ISD jobs, you can change the
$APT_CONFIG_FILE environment variable at the DataStage project level to
point to a single node configuration file as the default.

Important: Even when making a project-level change, you must still disable
and then re-enable the jobs in the ISD Console. However, recompilation of the
job is not necessary.

When setting the variable at the project level, be careful that the individual jobs
do not override the project-level default by defining a value locally.

Consider the performance implications. For complex jobs that process larger
data volumes, limiting the number of nodes in the configuration file might result in
reduced performance on a request-by-request basis. To ensure that overall
performance remains acceptable, you can increase the number of minimum job
instances that are available to service requests. See 3.8.2, “Scalability” on
page 93 for more information about multiple job instance servicing requests, as
determined at deployment time through the IS Console.

Chapter 3. Overview of IBM InfoSphere DataStage 99

 Proper monitoring must be in place to ensure that hardware resources are not
saturated and that enough spare capacity exists to accommodate additonal job
instances.

3.8.7 General considerations for ISD jobs

Consider when and how to use ISD with always-on jobs:
򐂰 Generally, the best job candidates for service publication with Information
Services Director have the following characteristics:
– Represent fine-grained data integration or data quality functions
– Are request/response-oriented
– Have small to medium payloads
– Are entirely idempotent (can be run over and over again in succession
without impact)
– Do little or no writing
– Have high watermark traffic requirements that are 300 requests per
second or less
򐂰 Job sequences cannot be published as services. If you have a business
requirement to be able to invoke a sequence of jobs via a service, design one
of the following jobs:
– Create a server job as a wrapper that uses job control logic only (no
stages) to invoke an actual DataStage job sequence.
– Deploy this wrapper job by using an asynchronous binding, such as JMS.

Synchronous binding: Using a synchronous binding (such as soap over

HTTP) can result in a timeout if the sequence runs for a long time.

3.8.8 Selecting jobs for publication through ISD

When selecting server or enterprise edition (EE) jobs for publication through ISD,
parallel jobs are the preference for the following reasons:
򐂰 Parallel jobs support pipeline and partitioning parallelism.
򐂰 Parallel jobs must be the preferred engine for any new development.
򐂰 They can scale on cluster and grid environments, and they are not restricted
to run on the head node only. They reduce the strain on the head node.
Therefore, it is less likely to require load balancing across multiple service
providers (DS engines).

100 InfoSphere DataStage for Enterprise XML Data Integration

򐂰 The transaction support enhancements that are described in 3.8.4, “ISD with
DTS” on page 96 and 3.8.5, “Information Services Director with connectors”
on page 98 overcome the previous limitations with parallel stages that made
DS server jobs necessary. The DTS and Database Connector stages in IS 8.5
provide for a similar functionality to a DS server combination of a Transformer
stage directly connected to an ODBC stage, for instance.

The following scenarios justify the publication of DS server jobs as ISD services:
򐂰 Publishing existing DS server jobs
򐂰 Invoking DS sequences:
– Sequences are in the Batch category and are outside of the scope of this
chapter.
– Sequences cannot be exposed as ISD services.

A DS server job can be created with a transformer that in turn invokes a DS

sequence with the utility sdk routine. The utility sdk routine is in the Transformer
category in the pull-down list of the Transformer. It is called UtilityRunJob.

Chapter 3. Overview of IBM InfoSphere DataStage 101

102 InfoSphere DataStage for Enterprise XML Data Integration
 4

Chapter 4. XML integration in

DataStage
XML technology is one of the most versatile methods for data interchange.
DataStage provides highly flexible integration capabilities and access to a host of
distributed enterprise data sources for industrial applications and shared
metadata repositories. Combining DataStage with XML offers opportunities to
develop powerful, industry-leading solutions that support enterprise initiatives,
such as customer relationship management (CRM) and decision support
systems (DSS).

DataStage was conceived as an extract, transform, and load (ETL) product in the
year 1996, the same year XML was adopted as a World Wide Web Consortium
(W3C) standard. Since then, it constantly evolved, adding new features and
capabilities and remaining a leading ETL tool.

Today, DataStage offers exceptional capabilities, such as the ability to connect to

many objects, including a wide variety of enterprise data sources (relational,
semi-relational, non-relational, and existing data):
򐂰 External applications, such as SAP
򐂰 Services, such as Web services
򐂰 Message queues, such as WebSphere MQ
򐂰 Other tools, such as InfoSphere Change Data Capture

© Copyright IBM Corp. 2012. All rights reserved. 103

 With the integration of QualityStage, which can be installed into DataStage as an
add-module, DataStage also provides in-stream data cleansing functionality. This
function eliminates the need for separate business processes for data
standardization and cleansing activities.

In addition, DataStage can expose user jobs as service-oriented architecture

(SOA) services, which allows external applications to use its services.
InfoSphere Master Data Management (MDM) Server and InfoSphere Metadata
Workbench (MWB) are examples of powerful applications that use the
DataStage SOA capabilities to provide value to the enterprises. The connectors
and adapters that are supported in DataStage can connect to external data
sources and also provide real-time data integration capabilities. The ability to
flow incoming and outgoing data by the use of XML Stage provides the ultimate
flexibility for handling large complex data and enhances the value of DataStage
as an ETL tool.

In this chapter, we review the available XML functionality in DataStage and how it
evolved over time. In the next several chapters, we explain the XML Stage, its
components, and how to work with it in detail. In the last two chapters, we
describe performance issues when handling large, complex XML documents and
designing XML-based solutions with DataStage.

104 InfoSphere DataStage for Enterprise XML Data Integration

4.1 DataStage components for handling XML data
DataStage relies on multiple types of components to access a wide variety of
enterprise applications and data sources, including XML. These components can
be divided into major groups, with the following terminology:
򐂰 Application programming interfaces (APIs)
APIs facilitate calling a DataStage job from an external application, such as a
Java program.
򐂰 Plug-Ins and Operators
Plug-ins and Operators are DataStage stages that are optimized for their
corresponding execution environments in Server and Parallel jobs.
򐂰 Stages
The stage components are the basic building blocks of a DataStage job.
Stages can be Built-in, Operators, Connectors, or Plug-Ins. Additional stages
that are not shipped with the product can be installed individually, developed
by the user, or included in Enterprise Packs.
򐂰 Packs and Enterprise Packs
Packs are special purpose components that consist of one or more stages,
often including specialized utilities for metadata import or other tasks.
Enterprise packs offer advanced features for accessing enterprise-wide tools,
such as Web services, InfoSphere Change Data Capture, and XML.
򐂰 Connectors
Connectors implement a common framework that can be used in DataStage,
Information Analyzer, and FastTrack, with a common connectivity framework
that also provides the capability to integrate with future tools.

We focus on the history of the implementation of XML technology, which evolved

with the standards of XML and with the connectivity component architecture that
is available to DataStage.

Chapter 4. XML integration in DataStage 105

4.2 The XML Pack
The following sections describe the various XML Packs.

4.2.1 XML Pack Version 1.0

The first generation of XML capabilities in DataStage, XML Pack 1.0, included
the XML Reader and XML Writer stages, which provided read and write access
to XML data. These stages were Built-in stages within Server jobs only. The
primary purpose of these components was to provide basic data conversion
between hierarchical XML and flat relational formats. XML Pack 1.0 relied on
Document Type Definition files (DTD files) for metadata import of document
structure, and the transformation capability was limited. As XML standards
evolved, XML schema led the industry as the method for the encapsulation and
validation of document structure. DataStage evolved with the industry, creating a
second generation of XML processing components, XML Pack 2.0. The two
packs coexisted for a few releases, but XML Pack 1.0 was eventually deprecated
in favor of the more feature-rich and capable newer version.

4.2.2 XML Pack Version 2.0

XML Pack Version 2.0 was implemented with the Plug-in architecture and was
available for use in both Server and Parallel jobs. Version 2.0 of the pack added
support for XML Schema Definition (XSD) schemas, schema validation, and
Extensible Stylesheet Language Transformation (XSLT) capabilities. It included
the XML Input, XML Output, and XML Transformer stages. It used industry
standard Xerces parser and Xalan XSLT engine technology, and the XPath
standard for identifying XML structure when parsing or composing documents.
XML Pack 2.0 is available in all currently supported DataStage releases that are
available at the time of writing this book.

Although the schema was used for validation at run time, and for metadata
import of the XPath and namespace information, the job developer needed to
manually design and configure the stage to ensure that the XML documents
were created to match the schema that was used. Additionally, when validation
was used, the schema must be accessible at run time either on disk or online
through a URL. The next generation of XML technology improves its
performance and eliminates these restrictions.

106 InfoSphere DataStage for Enterprise XML Data Integration

4.2.3 XML Pack Version 3.0
The third and current generation of XML capabilities in DataStage, XML Pack
3.0, includes a single XML Stage, which can be used as a source, target, or
transformation. Therefore, XML Pack 3.0 consolidates the functionality that
previously required three separate stages. The XML Stage adds internal schema
library management functionality that offers superior support for complex
schemas and advanced capabilities to simplify parsing and composing complex
XML documents.

In addition to the simplified single-stage design, the XML Stage is a Connector

stage, which is built on the common connectivity framework that was introduced
with Information Server. One of the exceptional features of the common
connectivity framework is “large object pass by” reference support. This feature
allows large chunks of text to be sent from one Connector stage to another with
minimal resource usage along the way. The connector framework passes only a
small reference value through the links. Then, the Connector stage retrieves the
large object directly from the source connector at run time when it is accessed.

The XML Stage uses a unique custom GUI interface, which is called the
Assembly Editor, that is designed to make the complex task of defining
hierarchical relationships and transformations as convenient and manageable as
possible. Chapter 6, “Introduction to XML Stage” on page 131 describes the
Assembly Editor in great detail. The XML Stage also can accept data from
multiple input links, and then establish a hierarchical relationship between the
two within the Assembly. For small data volumes, this capability can simplify job
design requirements and make the jobs easier to maintain. Chapter 10,
“DataStage performance and XML” on page 291 describes performance tuning
for the XML Stage, including how to manage large volumes of relational data and
large XML document sizes.

The most important feature of the XML Pack Version 3.0 is its tight integration
with XSD schema. After an XSD schema is imported by using the new Schema
Library Manager, the metadata from within the schema is saved locally. This
decoupling means that the schema files do not have to be accessible at run time.
This feature is important when deploying into restricted production environments.
The transformation and mapping functions directly link to the schema, so that the
job developer can focus on the business logic. The stage ensures that the
document conforms to the schema.

In the remaining chapters in this book, we fully explain the new XML Pack 3.0
functionality and performance tuning. We explain how to implement XML-based
solutions and migration strategies from XML Pack 2.0.

Chapter 4. XML integration in DataStage 107

4.3 Web Services Pack
The Web Services Pack facilitates real-time event-driven integration capabilities
in DataStage. It provides three stages for integrating Web Services (with SOAP
over HTTP protocol) into DataStage jobs:
򐂰 The WSClient (source) stage
򐂰 The WSClient (target) stage
򐂰 The WSTransformer (in-line transformation) stage

With these stages, you can supply input data to call a web service operation, and
retrieve the response as rows and columns. SOAP is an XML-based standard.
SOAP messages are XML documents that adhere to the SOAP schema.

The WSClient and WSTransformer stages can directly convert from relational to
SOAP XML, for simple structures. For more complex structures, with repeating
groups, the XML Stage is used in combination with the Web Service Pack stages
to compose or parse the SOAP message body or header, as needed. See 11.1,
“Integration with Web services” on page 330 for a complete solution to implement
a web service scenario with a complex structure in the SOAP message.

In 3.7, “Real-time applications” on page 79, we described real-time integration

with message queues and distributed transactions. DataStage contains an MQ
Connector stage for use as a source or target for IBM WebSphere MQ message
queues or Java Message Services (JMS) topics. The MQ connector can parse
relational data into rows. It also delivers the message contents as a single
payload value, which is useful for handling XML documents. See 11.3,
“Event-driven processing with IBM WebSphere MQ” on page 342 for a
description of using the MQ Connector in conjunction with the XML Stage to
process XML documents. This description includes important configuration steps
for the MQ Connector for job developers to implement this type of solution
successfully.

108 InfoSphere DataStage for Enterprise XML Data Integration

4.4 DataStage, Information Server, and XML metadata
In Chapter 3, “Overview of IBM InfoSphere DataStage” on page 45, we
introduced InfoSphere Information Server and described its unified metadata
infrastructure and the Metadata Workbench component. With Metadata
Workbench (MWB), you can explore and analyze technical, business, and
operational metadata about sources of data. With MWB, you can view the
complete lineage of data fields across jobs, applications, reports, or data
warehouses back to source systems, including the types of processing that were
used. Data that passes through the XML Stage in your jobs is included in this
data lineage sample. In Figure 4-1, we illustrate a sample Data Lineage report
that was generated by Metadata Workbench.

Figure 4-1 Sample Data Lineage report with XML Stage

InfoSphere Metadata Workbench supports both Data Lineage and Job Lineage
reports. Job lineage provides information about the flow of data through jobs and

Chapter 4. XML integration in DataStage 109

 physical data sources. It is a high-level view (business view) of the lineage. Data
Lineage provides much more granular information about the flow of data through
stages at the column level and includes XPath information for each of the
columns.

XPath represents one of the steps within the flow of data in either direction of a
selected metadata asset. It passes through a series of DataStage stages,
including an XML Stage and a sequence of DataStage jobs. It ends in databases
or other targets, such as business intelligence (BI) reports.

4.5 Exporting and importing DataStage repository

objects
DataStage includes several utilities for exporting and importing repository
objects, such as jobs, table definitions, routines, and shared containers to and
from physical files. Export and import are used for a number of purposes. With
them, you can migrate applications from one environment to another for
deployment purposes, for backing up applications, and for upgrades between
software releases. In addition to the commonly used DSX file format, XML format
is also available. XML is more human readable, making it a better choice if a
developer wants to compare the contents of export files before importing them.

The DataStage client includes Windows command-line utilities for automating the
export/import processes, when needed. Using the utilities in conjunction with the
XML export format was one method of integrating with external source control
systems. The built-in source control integration that is now available with
Information Server significantly replaces this manual XML-based method.

The Information Server Manager tool and its command-line version istool
provide the source control integration features and robust packaging options to
give the developers greater control of how objects are backed up and deployed.
The .isx file that is created by Information Server Manager is actually a
compressed file/jar format. The files that are inside that represent the contents
are actually XML files. Of particular interest is the manifest file, which is located
within the export file at ../META-INF/IS-MANIFEST.MF. This file is an XML file, and
it contains an index of all the repository objects that are contained within the
export file. Figure 4-2 on page 111 contains an excerpt from an IS-MANIFEST.MF
file.

110 InfoSphere DataStage for Enterprise XML Data Integration

Figure 4-2 Excerpt from IS-MANIFEST.MF file within an ISX export

Each object that is contained in the export file includes an <entries> element.
The developer can extract the IS-MANIFEST.MF file and use this index to identify
the contents of the file as needed without running the file through the tool. For
information about the Information Server Manager, see the links to the online
product documentation in , “Related publications” on page 375.

Chapter 4. XML integration in DataStage 111

112 InfoSphere DataStage for Enterprise XML Data Integration
 5

Chapter 5. Schema Library Manager

In this chapter, we describe how to use the Schema Library Manager for
importing XML schemas into the metadata repository to use in the XML Stage
when designing the DataStage job. We also describe how these schemas are
grouped into libraries that help in the management of many schemas inside a
single project, and the benefits of XML schemas inside a common repository.
The first step in configuring the XML Stage is to use the schemas in the
repository. In the subsequent chapters, we describe the XML Stage in detail.

The metadata repository stores imported metadata, project configurations,

reports, and results for all components of IBM InfoSphere Information Server.

You can access the information in the metadata repository from any of the suite
components, which means that metadata can be shared by multiple suite
components. For example, after importing metadata into the repository, an
authorized user can then use the same metadata in projects in any of the suite
components, such as IBM InfoSphere FastTrack, InfoSphere DataStage, or
InfoSphere QualityStage.

The XML Stage can be used to read, write, and transform XML data. To read,
write, and transform XML files, the XML schema is required. Without a valid
schema, you cannot use the XML Stage. The XML schema defines the structure
of the XML file. That is, it defines the metadata of the XML file. The XML schema
is analogous to the columns that are defined in the Sequential File stage. The
columns in the Sequential File stage describe the data that is read by the

© Copyright IBM Corp. 2012. All rights reserved. 113

 Sequential File stage. Similarly, the XML schema defines the data that is present
in the XML file.

You must store the XML schema in the metadata repository to use it in the XML
Stage while designing the DataStage job. You use the Schema Library Manager
to import the XML schema into the metadata repository. You access the Schema
Library Manager from the DataStage Designer Client by selecting Import 
Schema Library Manager, as shown in Figure 5-1.

Figure 5-1 The Schema Library Manager in the Import menu

5.1 Creating a library

When you open the Schema Library Manager, the Schema Library Manager
window opens as shown in Figure 5-2 on page 115. In the Schema Library
Manager window, the schema is imported into a library. The library is used to
organize schemas into a single group. Therefore, to import an XML schema into
the Schema Library Manager, you must first create the library.

114 InfoSphere DataStage for Enterprise XML Data Integration

Figure 5-2 Schema Library Manager window

Use the following steps to create a library in the Schema Library Manager:
1. In the Schema Library Manager window, click New Library. The New
Contract Library window opens, as shown in Figure 5-3.

Figure 5-3 New Contract Library window

2. In the New Contract Library window, in the Name text box, enter the name of
the library, which can be any valid name that begins with an underscore,
letter, or colon. Enter the library name as Test, as shown in Figure 5-4 on
page 116.

Chapter 5. Schema Library Manager 115

 Figure 5-4 Name text box in the New Contract Library window

3. In the Description text box, enter a short optional description of the library.
Enter This is a test library in the Description field, as shown in
Figure 5-5. Click OK.

Figure 5-5 Description text box in the New Contract Library window

A library with the name Test is created, as shown in Figure 5-6 on page 117.

116 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 5-6 The Test library in the Schema Library Manager

Multiple libraries can be created in the Schema Library Manager, but each library
name must be unique. Therefore, two libraries cannot have the same name. The
names are case-sensitive; therefore, the names Test and test are unique.

A library can be accessed through all DataStage projects; therefore, our library
Test can be accessed through all DataStage projects.

5.2 Importing an XML schema into a library

After creating a library, you can import an XML schema into it. Use the following
steps to import the schema employee.xsd into the library Test:
1. Select the library Test; the Resource View section becomes active, as shown
in Figure 5-7 on page 118.

Chapter 5. Schema Library Manager 117

 Figure 5-7 The Resource View section in the Schema Library Manager

2. Click Import New Resource. You can browse through the folders in the client
machine. Select the schema employee.xsd and click Open. The
employee.xsd file is imported. You are notified through the window, as shown
in Figure 5-8. Click Close.

Figure 5-8 The notification window while importing a schema into the library

The employee.xsd is imported into the library Test, as shown in Figure 5-9 on
page 119.

118 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 5-9 Schema file is imported into the Schema Library Manager

Two schemas with the same name cannot be imported into the same library. But,
two libraries can hold the same schema. Therefore, you can have the same set of
schemas in two separate libraries.

5.3 Categorizing the libraries

A library in the Schema Library Manager can be added to a category. A category
is used to organize multiple libraries into a single group.

A library can be added to a category in two ways: while creating the library and
after creating the library. Follow these steps:
1. When you create a library, you can enter a name in the Category text box, as
shown in Figure 5-10 on page 120.

Chapter 5. Schema Library Manager 119

 Figure 5-10 The Category field in the New Contract Library window

2. If the library is created without a category, the library can still be categorized.
To categorize an existing library, select the library. The details of the library
are listed, as shown in Figure 5-11. You can enter a category name in the
Category text box.

Figure 5-11 The Category option for an existing library

When you add the library Test to the Category Cat1, you create a folder called
Cat1. The Test library is in the folder, as shown in Figure 5-12 on page 121.

120 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 5-12 The library Test in the category Cat1

You cannot have two categories with the same name; each category must have a
unique name. The names are case-sensitive; therefore, the names Cat1 and cat1
are unique. Two categories cannot have the same library. For example, the
library Test cannot exist in two separate categories.

5.4 Viewing the XML schema structure

When an XML schema is imported into the Schema Library Manager, it is
translated into a simplified model. You can see this simplified structure of the
XML schema in the Schema Library Manager. For example, the schema
employee.xsd is imported into a library Test. To see the structure of the schema,
double-click the library Test, or select the library Test and click Open. The Types
tab opens, as shown in Figure 5-13 on page 122.

Chapter 5. Schema Library Manager 121

 Figure 5-13 The Types tab in the Schema Library Manager

The global elements of the schema shows on the Types page. In employee.xsd,
there are two global elements, as shown in Figure 5-13. The elements do not
have a namespace; therefore, the namespace is not shown.

When you click the global element employee, you can see its structure on the
right side of the window, as shown in Figure 5-14 on page 123.

122 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 5-14 Structure of the element employee

5.5 Importing related schema files

Multiple schemas can relate to one another; therefore, one schema can
reference the elements and complexTypes that are used in another schema.
Schema files can relate to each other by an INCLUDE or IMPORT statement.
Reference other schema files through the Schema Location (INCLUDE) or the
Namespace (IMPORT) of the schema file. Such references must be resolved
within a single schema library. Also, you need to include all files that are
referenced by a schema file in the same library.

5.5.1 IMPORT statement

Two schemas can reference one another by using the IMPORT statement. The
IMPORT statement uses the namespace as the ID of the schema file. Because
both the schemas related to one another, both the schemas must be imported
into the same library in the Schema Library Manager. The Schema Library
Manager automatically relates both the schemas to create the simplified schema
structure. No user intervention is required. If you import only one schema, the
library shows an error.

Chapter 5. Schema Library Manager 123

 We show an example of the two schemas that relate to each other by using the
IMPORT statement in Example 5-1.

Example 5-1 Schemas that use the IMPORT statement

--<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
targetNamespace="http://ibm.com/infosphere/xml/Organization"
xmlns:prn="http://ibm.com/infosphere/xml/Employee">
<xs:import namespace="http://ibm.com/infosphere/xml/Employee"
schemaLocation="import_schema2.xsd"/>
<xs:element name="departments">
<xs:complexType>
<xs:sequence>
<xs:element name="department" minOccurs="1"
maxOccurs="unbounded">
<xs:complexType>
<xs:sequence>
<xs:element name="departmentId" type="xs:string"/>
<xs:element name="departmentName" type="xs:string"/>
<xs:element ref="prn:employees"/>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:schema>
_________________________________________________________
<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
targetNamespace="http://ibm.com/infosphere/xml/Employee">
<xs:import namespace="http://ibm.com/infosphere/xml/Organization"
schemaLocation="import_schema1.xsd"/>
<xs:element name="employees">
<xs:complexType>
<xs:sequence>
<xs:element name="name" type="xs:string"/>
<xs:element name="dob" type="xs:date"/>
<xs:element name="title" type="xs:string"/>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:schema>

124 InfoSphere DataStage for Enterprise XML Data Integration

5.5.2 INCLUDE statement
The INCLUDE statement uses the location of the file that relies on a physical
location. When the schema files are imported into the Schema Library Manager,
the location attribute of each file defaults to the file name. However, the files are
commonly referenced by more than their file names. For example, Schema A can
reference Schema B by using a relative directory structure:
<xs:include schemaLocation="../common/basic.xsd">

Or, it can even reference a URL to a web-hosted file:

<include schemaLocation ="http://www.example.com/schemas/address.xsd">

In these cases, you must modify the file location attribute, as shown in
Figure 5-15, of the referenced file in the Schema Library Manager to the location
used by the INCLUDE statements.

Figure 5-15 The File Location for an imported schema file

The following steps demonstrate how to import two schema files

(include_schema1.xsd and include_schema2.xsd) with an include dependency
between them:
1. Import the schema include_schema1.xsd into the library Test.

Chapter 5. Schema Library Manager 125

 2. The library shows errors. The error message indicates that the type definition
for some of the elements cannot be found, because the definitions are in the
schema include_schema2.xsd. Import include_schema2.xsd into the same
library.
3. Even after importing include_schema2.xsd, errors are still seen in the library.
The location attribute in the include_schema1.xsd schema points to the actual
physical location of the file, which is
http://ibm.com/definitions/definition.xsd. However, in the Schema
Library Manager, the location attribute defaults to the schema name, for
example, include_schema2.xsd.
4. To resolve the errors, change the file location of the schema
include_schema2.xsd, as shown in the following steps:
a. Click the include_schema2.xsd schema in the Resources View window.
b. In the File Location field, replace include_schema2.xsd with
http://ibm.com/definitions/definition.xsd, as shown in Figure 5-16.

Figure 5-16 Changing file location to match the URL in the INCLUDE statement

c. When you change the file location name, the library turns green, which
indicates that the library is valid.

126 InfoSphere DataStage for Enterprise XML Data Integration

5.5.3 Importing multiple schemas
Multiple XML files can be imported into a single library by using two methods:
1. When browsing for an XML schema, multiple files can be selected, as shown
in Figure 5-17.

Figure 5-17 Selecting multiple files in the browse window

2. The Schema Library Manager can import an entire compressed file that
contains many XSD files. When a compressed file is imported, the Schema
Library Manager imports all the files in the compressed file and sets the
location attribute to the relative directory from the root of the compressed file.
This feature can save the tedious work of importing all the files one at a time
and updating their locations.

5.6 Errors while importing schemas

After importing a schema into a library, the library is automatically validated to
determine whether the schema contains any errors. If the validation is
successful, the library contains all the element and type definitions from the

Chapter 5. Schema Library Manager 127

 schemas. If the library is invalid, you are notified about the errors. When an
invalid schema is imported into the Schema Library Manager, the library shows a
red cross to the left of it, as shown in Figure 5-18.

Figure 5-18 The library shows errors

To view the list of errors, click Validate. Whenever you modify a schema, delete a
schema from the library, or add a schema to the library, the library is
automatically revalidated.

Libraries can show the following errors:

򐂰 The library shows errors if the XML schema is invalid. For example, the XML
schema is missing a closing tag for one of the element definitions.
򐂰 Two schema files have the same global elements. When both the schemas
are imported into a single library, the library becomes invalid and shows the
error “A schema cannot contain two global components with the same
name”. Therefore, two schemas with the same elements cannot be imported
into a single library.
򐂰 Consider two schemas that relate to each other with the IMPORT statement.
If a single schema is imported into the library, the library becomes invalid. For
the library to be valid, both schemas must be imported into a single library.
򐂰 Two schemas can relate to each other with the INCLUDE statement. When
you import both schemas into the same library, the library can still be invalid.
To make the library valid, the file name URL must be modified, as described in
5.5.2, “INCLUDE statement” on page 125.

128 InfoSphere DataStage for Enterprise XML Data Integration

5.7 Accessing the Schema Library Manager from the
XML Stage
The XML Stage consists of the Assembly Editor. You use the Assembly Editor to
design the XML data. You can add steps to read, write, and transform XML data.
You access the Schema Library Manager from the Assembly Editor by using the
Libraries tab, as shown in Figure 5-19. This tab allows the user to import new
schemas without needing to close the Assembly Editor.

Figure 5-19 Accessing the Schema Library Manager through the Assembly Editor

In this chapter, we described how to use the Schema Library Manager to import
XML schemas into the metadata repository so that it can be used in the XML
Stage while designing the DataStage job. We also described how these schemas
are grouped into libraries that help in the management of many schemas inside a
single project. We explained the benefits of using XML schemas inside a
common repository. Think of putting the schemas in the repository as a first step
in configuring the XML Stage. In the subsequent chapters, we describe the XML
Stage in detail.

Chapter 5. Schema Library Manager 129

130 InfoSphere DataStage for Enterprise XML Data Integration
 6

Chapter 6. Introduction to XML Stage

The XML Stage was introduced in InfoSphere DataStage 8.5 (all previous
solutions are called the XML Pack). The XML Stage can support hierarchical
metadata and perform hierarchical transformations. It is based on a unique
state-of-the-art technology to parse and compose any complex XML structure
from and to a relational form or a different hierarchical form. Earlier versions of
InfoSphere DataStage parsed, composed, and transformed XML data by using
three stages, the XML Input, XML Output, and XML Transformer stages. But, the
XML Stage can perform all these operations in a single stage.

Now, InfoSphere DataStage supports two XML solutions: the XML Pack and the
XML Stage. The XML Pack, which includes the XML Input, XML Output, and
XML Transformer stages, is useful to perform simple transformations that do not
involve a large amount of data. The XML Stage is the best choice for an XML
solution, because it can perform complex transformations on a large amount of
data.

The XML Stage differs from other known XML tools because of its powerful
execution, which can process any file size, with limited memory and in parallel.
The XML Stage has unique features, such as the ability to control and configure
the level of validation that is performed on the input XML data. Language skills,
such as Extensible Stylesheet Language Transformation (XSLT) or Xquery, are
not required to use XML Stage.

© Copyright IBM Corp. 2012. All rights reserved. 131

6.1 The XML Stage
The XML Stage is in the Real Time section of the palette in the DataStage
Designer and contains all the components that are required to process any XML
document.

The XML Stage consists of the Stage Editor and the Assembly Editor. The Stage
Editor is the first window that opens when you click the stage, as shown in
Figure 6-1. The runtime properties of the stage are defined in the Stage Editor.

Figure 6-1 The Stage Editor

Click Edit Assembly on the Stage Editor to open the Assembly Editor. You
design and create an assembly in the Assembly Editor. An assembly consists of a
series of steps that enriches and transforms the XML data, as depicted in
Figure 6-2 on page 133.

132 InfoSphere DataStage for Enterprise XML Data Integration

The assembly is analogous to a car factory. In a car factory, there are multiple
stations and in each station a new part is added to complete the car. Each station
performs only one task, and the car moves along all the stations until it is
complete. Similarly, each step in the assembly acts on the incoming data until all
the transformations are finished and the assembly is complete.

Aggregate

Aggregate
Output
Output

Output
Input

Input

Input

Join
Step 1 Step 2 Step n

Figure 6-2 The Assembly process

When you open the Assembly Editor, as shown in Figure 6-3, by default, it
includes the Input and the Output steps.

Figure 6-3 The Assembly Editor

Chapter 6. Introduction to XML Stage 133

 The XML Stage contains multiple steps, and each step has a specific
functionality. These steps are used together to design a solution. The steps are in
the Palette in the Assembly Editor. The steps are added between the Input and
the Output steps, to parse, compose, and transform XML data. The XML Stage is
similar to a DataStage job. In the DataStage job, multiple stages are used to
process the incoming data. Similarly, in the XML Stage, multiple steps are used
to transform the data.

6.2 Schema representation in the Assembly Editor

The XML Stage is a schema-driven stage. The XML schema is required to parse,
compose, or transform an XML file. You cannot start designing the assembly
without a valid schema. The schema, that is, the xsd file, needs to be imported
into the Schema Library Manager so that the schema metadata can be accessed
by the steps within the assembly.

When an XML schema is imported into the Schema Library Manager, it is

translated into a simplified model. This simplified model is used by all the
transformation steps that are available in the assembly. In this model, the
primitive data types are the same primitive types that the XML schema defines.
However, when describing complex data, a few of the XML schema concepts are
simplified, although all the information from the XML schema is still preserved.

It is important that you understand the schema representation in the assembly.

An understanding of the schema representation helps you to understand deeper
concepts that are required while designing the assembly. The schema
representation is the foundation on top of which the actual building (assembly) is
built.

In the following sections, we provide a detailed description of the schema

representation in the assembly.

6.2.1 The schema structure

The schema is depicted with a tree structure in the assembly. In the tree
structure, the child elements are placed underneath the parent elements. As
shown in Figure 6-4 on page 135, department is the parent of employee. The
elements employee_name and age are the children of the element employee. By
using the plus sign (+) or the minus sign (-) next to each parent element, the child
elements can be hidden or unhidden from view.

134 InfoSphere DataStage for Enterprise XML Data Integration

 <?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name=" department">
<xs:complexType>
<xs:sequence>
<xs:element name="employee"> department
<xs:complexType>
<xs:sequence> employee
<xs:element name="employee_name" type="xs:string"/>
<xs:element name="age" type="xs:int"/> employee_name
</xs:sequence>
</xs:complexType> age
</xs:element>
</xs:sequence> The schema representation
</xs:complexType> in the assembly.
</xs:element>
</xs:schema>
The schema file

Figure 6-4 Schema representation in the Assembly

This pictorial representation of the schema in the assembly helps you to see the
structure of the incoming XML and clearly establish the relationship between
various elements within the schema.

6.2.2 Primitive data type representation

A primitive-type element in the XML schema is called an item in the assembly.
Therefore, an item in the assembly corresponds to a single primitive data type in
the XML schema. In Figure 6-5, we show the representation of a few elements
within the assembly. Each element has a unique symbol before it that is based on
the data type of the element.

<xs:element name="book_title" type="xs:string"/>

<xs:element name="date_of_publish" type="xs:date"/>

<xs:element name="code" type="xs:ID"/>

<xs:element name="author" type="xs:string"/>

<xs:element name="cost" type="xs:decimal"/>

Representation in
Definition in the XSD the Schem a

Figure 6-5 Primitive types representation in the Assembly

Chapter 6. Introduction to XML Stage 135

6.2.3 Attribute representation
The attributes in the schema are represented by an item with a leading at sign
(@). As shown by the representations in Figure 6-6, the attribute type is
represented as an item with the name @type. Therefore, from the
representations in the assembly, name is an element, and type, age, and breed
are attributes.

<xs:element name="pet">
<xs:complexType>
<xs:sequence>
<xs:element name="name" type="xs:string"/>
</xs:sequence>
<xs:attribute name="type" type="xs:string"/>
<xs:attribute name="age" type="xs:int"/>
<xs:attribute name="breed" type="xs:string"/>
</xs:complexType>
</xs:element>

Definition in the XSD Representation in

the Schema

Figure 6-6 Attribute representation in the Assembly

6.2.4 Representation of repeating elements (LIST)

In the XML document, elements might occur more than one time. In the XML
schema, these elements show that maxOccurs is defined to a value greater than
1, as shown in Example 6-1.

Example 6-1 Repeating elements

<xs:element name="employee_name" type="xs:string" maxOccurs="3">

These repeating elements of the schema are represented by a list in the

assembly. As shown in Figure 6-7 on page 137, the element "phoneNumber" can
occur multiple times in the XML file. In the assembly, this element is depicted by
a blue icon, which represents it to be a list. The list in the assembly holds the
actual number of occurrences of the element. The item text() under the list holds
the actual data for the element.

136 InfoSphere DataStage for Enterprise XML Data Integration

 <xs:element name="company">
<xs:complexType>
<xs:sequence>
<xs:element name="name" type="xs:string"/>
<xs:element name="CEO" type="xs:string"/>
<xs:element name="headquarter" type="xs:string"/>
<xs:element name="year_of_establishment" type="xs:string"/>
<xs:element name="founder" type="xs:string"/>
<xs:element name="phoneNumber" type="xs:string" maxOccurs="unbounded"/>
</xs:sequence>
</xs:complexType>
</xs:element>

Definition in the XSD

Representation
in the Schema
Figure 6-7 Repeating elements representation in the Assembly

NMTOKENS-type representation
The XML schema can have elements of type NMTOKENS. In these elements,
maxOccurs might not be set to a value greater than 1 in the XML schema. In the
assembly, these elements are represented by a list, because the data type
NMTOKENS represents an array of data separated by spaces. Because multiple
data values are held by a single element, the assembly represents this element
as a list. As shown in Figure 6-8, the element "codes" is represented by a list with
the same name. The actual value of the element is held by the item text().

<xs:element name="appliance">
<xs:complexType>
<xs:sequence>
<xs:element name="codes" type="xs:NMTOKENS"/>
</xs:sequence>
</xs:complexType>
</xs:element>
Representation in
Definition in the XSD
the Schema

Figure 6-8 NMTOKENS-type representation in the Assembly

In the XML schema, when an element of type NMTOKENS shows maxOccurs

set to a value greater than 1, the element is represented by two successive lists.
One list is represented by the name of the element, and the other list is
represented by anonymousList (Figure 6-9 on page 138).

Chapter 6. Introduction to XML Stage 137

 <xs:element name="appliance">
<xs:complexType>
<xs:sequence>
<xs:element name="codes" type="xs:NMTOKENS" maxOccurs="unbounded"/>
</xs:sequence>
</xs:complexType>
</xs:element>

Definition in the XSD Representation in

the Schema

Figure 6-9 NMTOKENS with maxOccurs greater than 1 representation in the Assembly

xs:List representation
An element of type xs:list in the schema is also represented by a list in the
assembly, as shown in Figure 6-10.

<xs:element name="appliance">
<xs:simpleType>
<xs:list itemType="xs:int"/>
</xs:simpleType>
</xs:element>

Definition in the XSD Representation in

the Schema

Figure 6-10 List type representation in the Assembly

6.2.5 DataStage links representation

DataStage input and output links are also represented as a list in the assembly,
because each link can carry multiple rows of data during run time. An example of
an input link is shown in Figure 6-11 on page 139. The input link DSLink1 links to
the XML Stage. The corresponding representation in the Assembly Editor is a list
with the following items: name, age, dob, and hireDate. These items are defined
columns in the DataStage link DSLink1.

138 InfoSphere DataStage for Enterprise XML Data Integration

 XML Stage with input link DSLink1

Representation in the Schema

Figure 6-11 Input links representation in the Assembly

6.2.6 Complex-type representation

In the assembly, a group is a node in the hierarchy whose only purpose is to
create a structure that categorizes or encapsulates information. It does not
describe a dimension or value. A group can correspond to a complex element
with the type: sequence or choice. In Figure 6-12, the element colony is a
complex element with type sequence. Therefore, it is represented as a group
within the assembly. A group is depicted by a brown icon with a subscript of G.

<xs:element name="colony">
<xs:complexType>
<xs:sequence>
<xs:element name="house" maxOccurs="unbounded">
<xs:complexType>
<xs:sequence>
<xs:element name="number" type="xs:integer"/>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType> Representation in
</xs:element> the Schema

Definition in the XSD

Figure 6-12 Complex-type representation in the Assembly

Specific schema definitions are represented in certain ways in the assembly. We

explain several schema representations in the following list:
򐂰 When a complex element is defined with a choice, an item
@@choiceDiscriminator is added to the schema representation in the
assembly. This item holds the value of the branch that was selected at run

Chapter 6. Introduction to XML Stage 139

 time. As shown in Figure 6-13, the element pet is a choice. Therefore, the
item @@choiceDiscriminator is added as a child of the group pet.

<xs:element name="pet">
<xs:complexType>
<xs:choice>
<xs:element name="dog" type="xs:string"/>
<xs:element name="cat" type="xs:string"/>
<xs:element name="parrot" type="xs:string"/>
<xs:element name="cow" type="xs:string"/>
</xs:choice>
</xs:complexType>
</xs:element>

Definition in the XSD Representation in

the Schema

Figure 6-13 Complex choice type representation in the Assembly

򐂰 For an element with mixed content, two new items are added in the assembly:
@@mixedContent and @@leadingMixedContent:
– @@leadingMixedContent captures the mixed content at the start of the
element.
– @@mixedContent captures the data that follows the element. As shown in
Figure 6-14, the complex element "location" is of mixed type.

<xs:element name="employee">
<xs:complexType>
<xs:sequence>
<xs:element name="name" type="xs:string"/>
<xs:element name="age" type="xs:int"/>
<xs:element name="dob" type="xs:date"/>
<xs:element name="location">
<xs:complexType mixed="true">
<xs:sequence>
<xs:element name="address" type="xs:string"/>
</xs:sequence>
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType>
</xs:element>
Representation in
Definition in the XSD the Schema

Figure 6-14 Elements with mixed content representation in the Assembly

6.2.7 Representation of recursive elements

A recursive element is represented with the infinity symbol as its suffix. An
element is called a recursive element if it refers to itself in the XML schema. In
Figure 6-15 on page 141, the element address is a recursive element, because
its child element also references address.

140 InfoSphere DataStage for Enterprise XML Data Integration

 <?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="employees">
<xs:complexType>
<xs:sequence>
<xs:element ref="address" maxOccurs="unbounded"/>
</xs:sequence>
</xs:complexType>
</xs:element>
<xs:element name="address">
<xs:complexType>
<xs:sequence>
<xs:element ref="address" maxOccurs="unbounded"/>
<xs:element name="emp_id" type="xs:string"/>
</xs:sequence> Representation in
</xs:complexType> the Schema
</xs:element>
</xs:schema>
Definition in the XSD

Figure 6-15 Recursive element representation in the Assembly

6.2.8 Representation of derived types

When an element is derived from another element, the item @@type is added to
the schema representation in the assembly. As shown in Figure 6-16, the
element Information is of type Person. Because Employee is derived from
Person, the element Information can also be of type Employee. Therefore, the
Item @@type is added under Information, so that at run time it can decide which
structure to use.

<?xml ver sion="1.0" encoding="UTF-8"?>

<xs:schem a xmlns:xs=" http://w ww .w 3. org /2001/XML Schema">
<xs:co mpl exType name="Per so n">
<xs:seq uence>
<xs:elem ent name="firstName" type="xs:string" / >
<xs:elem ent name="lastNam e" type="xs:string " />
</xs:sequence>
</xs:com plexType>
<xs:co mpl exType name="Employee">
<xs:com plexC onten t>
<xs:extensi on b ase="Person">
<xs:sequence>
<xs:elem ent name="emp_id " type="xs:strin g" />
</xs:sequence>
</ xs:extension>
</xs:co mplexContent>
</xs:comp lexT ype>
<xs:el ement name="top">
<xs:com plexT yp e>
<xs:sequence>
<xs:element nam e="I nform atio n" m axO ccurs="unb ounded" type="Per so n" />
</ xs:seq uence>
</xs:co mplexType>
</xs:element> Representation in
</xs:schema> the Schema

Definition in the XSD

Figure 6-16 Element derivation representation in the Assembly

Chapter 6. Introduction to XML Stage 141

6.2.9 Representation of an element with all optional elements
An element in the XML schema can have all optional child elements. When an
element is optional, minOccurs is set to "0" in the XML schema. In the assembly,
the parent element with all optional child elements has an extra item called
@@isPresent added to it. As shown in Figure 6-17, the element employee has all
optional child elements. Therefore, the item @@isPresent is added as a child of
the element employee.

<?xml version="1.0" encoding ="UTF-8"?>

<xs:schem a xm lns:xs="http://ww w .w 3. org/ 2001/XMLSchema">
<xs:elem ent name="Department">
<xs:com plexT yp e>
<xs:sequ ence>
<xs:element nam e="em ployee" >
<xs:co mpl exType>
<xs:sequence>
<xs:elem ent nam e="n am e" typ e="xs:string" m inO ccurs="0"/>
<xs:elem ent nam e="age" type="xs:in t" minOccurs="0" />
<xs:elem ent nam e="d ob" type="xs:date" minOccurs="0"/ >
<xs:elem ent nam e="em p_id" type="xs:str ing" minOccurs="0"/>
</xs:sequence>
</xs:co mplexType>
</xs:elem ent>
</xs:sequence>
</xs:co mpl exType>
</ xs:element>
</xs:schema>
Representation in
Definition in the XSD the Schema

Figure 6-17 Optional element representation in the Assembly

When an XML schema is imported into the library, it is translated into a simplified
model. All the Transformation steps use these simplified concepts. For example,
in the Aggregate step, aggregating an attribute of type double is the same as
aggregating an element of type double. The only steps that are sensitive to the
original XML schema concepts are the XML Parser and Composer steps. Due to
the ability to preserve the information in the XML schema, these steps can
recreate the XML schema concepts and adhere to them, as needed.

6.3 The assembly model

The assembly is used to describe the hierarchical data that flows through the
assembly steps. The assembly computational model differs from the DataStage
job computational model in the following ways:
򐂰 The DataStage job design is well-suited for relational data; the assembly is
well-suited for hierarchical data.
򐂰 In a DataStage job, each link carries a relation. Therefore, if a stage
transforms more than one relation, two input links need to be attached to the

142 InfoSphere DataStage for Enterprise XML Data Integration

 stage, and the output behavior is similar. However, in the assembly, a single
input can carry multiple lists; therefore, all steps have a single input and a
single output, each of which contains multiple lists.
򐂰 The assembly steps, similar to the DataStage stages, always iterate input
lists. However, unlike DataStage stages, steps iterate lists that exist in any
level of the hierarchical input and produce output that is contained in the input
list that they iterate. As shown in Figure 6-18, in the DataStage job, Stage3
can perform transformations only on the output of Stage2 and not on the
output of Stage1. But in the assembly, Step3 can process the data output of
Step1 and Step2. The output of Step1 is not fed to Step3. Therefore, each
step can access the output of any other previous step.

Stage1 Stage2 Stage3

Step1 Step2 Step3

Figure 6-18 Comparison between DataStage jobs and Assembly computational model

The input and output steps of the assembly have a distinguished role. The Input
step transforms relational data into hierarchical data, and the Output step
transforms hierarchical data to relational data. Each input/output link to the XML
Stage turns into a list in the assembly. Each column turns into an item of the list,
as shown in Figure 6-11 on page 139.

In the assembly, the entire input tree is rooted in a list named top. Therefore, top
is the parent of all lists and items in the assembly. Consider that the XML Stage
has two input links: DSLink1 and DSLink2. As shown in Figure 6-19 on page 144,
these links turn into a list in the assembly, and both links are present under the
list top.

Chapter 6. Introduction to XML Stage 143

 DataStage input links
Representation in
the Schema
Figure 6-19 Input step representation in the Assembly

Similarly, when multiple steps are used, the input and output of each step is
present under the list top.

When input links exist to the XML Stage, all the input lists in the assembly are
categorized under the group InputLinks. This group is the child of the list top, as
shown in Figure 6-19. Even when no input links exist to the XML Stage, the
group InputLinks is still present under top, but it does not contain any child lists
or items.

The steps in the assembly pass the entire input to the output along with an
enrichment branch that contains the result of the step. The output of a step is
always contained within the top-level list through which the current step iterates.
The Assembly Editor presents for each step its input and output trees. The
Assembly Editor highlights the enrichment branch that the current step
computed, so that the enrichment branch can be distinguished from the original
input of the step. As shown in Figure 6-20 on page 145, the output of the Sort
step is highlighted by graying out the area.

144 InfoSphere DataStage for Enterprise XML Data Integration

Figure 6-20 Output step representation in the Assembly

This powerful assembly data model simplifies the transformation description into
a series of steps rather than a direct graph.

Chapter 6. Introduction to XML Stage 145

146 InfoSphere DataStage for Enterprise XML Data Integration
 7

Chapter 7. Consuming XML documents

In the scenarios in this chapter, we reference a fictitious healthcare company that
we call the RedbookXML Healthcare Company. The company, which was
established in the late 1990s, receives insurance claims from residents in and
around San Jose, California. At the company, a web application is used to
capture the claims from their customers, and this data is stored in XML files. At
the end of every month, a report is created that summarizes the data for the total
number of claims in that month. The application that creates these reports
requires the data to be in a database. Figure 7-1 shows this process.

The
Theclaim
claimdata
data
A web application gets
gets converted
convertedto to
captures the XML
XMLfiles
files
customer claim
Customers

Reports are then The XML file is parsed

created with the data and the data then stored
from the database in the database

Figure 7-1 Healthcare company claims process

© Copyright IBM Corp. 2012. All rights reserved. 147

7.1 XML Parser step
Because the company requires data that is resident in a database to generate
the monthly reports, the company needed a way to extract the data from the XML
files and to store it in the database. The company already uses InfoSphere
DataStage for processing and cleansing its customer information. Therefore, the
company wants to use the XML stages that are available with DataStage to
process the XML claim files.

First, the company parses the XML file with either the XML Input stage or the
XML Stage within the DataStage Designer. The company selects the XML Stage,
because it can process large XML files easily. Within the XML Stage, the Parser
step is used, and it is added to the Assembly Editor. Follow these steps:
1. Open Assembly Editor and click Palette in the Assembly Outline bar. The
steps are shown in the Palette, as shown in Figure 7-2.

Figure 7-2 The Palette in the Assembly Editor

148 InfoSphere DataStage for Enterprise XML Data Integration

 2. Double-click XML_Parser Step in the palette. The XML_Parser Step is added
between the Input Step and the Output Step, as shown in Figure 7-3.

Figure 7-3 XML_Parser Step added to the Assembly Editor

The company stores all the XML files in two central repositories: the database
and one of its internal servers. The Parser step must read the XML files from
these locations to parse the XML data. In the XML Parser step, the method of
reading the XML data must be specified in the XML Source tab. A single Parser
step cannot read the XML data from multiple sources at the same time.

7.1.1 Reading XML data from a database

In the fictitious healthcare company, we use DB2 for the database storage. The
XML files are stored in a single DB2 pureXML column in the database. In a
pureXML column, the XML files can be stored in their native hierarchical format
without needing to shred the data into relational format. Applications can access
the XML data in this column. In the DataStage Designer canvas, you can use the
DB2 Connector stage to read the XML files from the database. The XML files are
stored in a single column, claim_XML, in the database. The DB2 Connector can
access the appropriate database table to read this particular column. The output
of the DB2 Connector stage is given to the XML Stage, as shown in Figure 7-4.

Chapter 7. Consuming XML documents 149

 Figure 7-4 The DataStage job to read XML files from the database column

The Input step in the Assembly Editor describes the metadata of a relational
structure. The step becomes active when an input link is provided to the XML
Stage. You can view the columns that are added to the preceding stage of the
XML Stage here. You can modify, remove, or add the columns. When the input is
provided from the DB2 Connector to the XML Stage, the Input Step becomes
active and the DB2 column, claim_XML, is displayed, as shown in Figure 7-5.

Figure 7-5 The Input Step in the Assembly Editor

To read the data from the database column, you must configure the XML Source
tab of the Parser Step. You must use the String set option in the Parser Step to
read the XML data from the database column claim_XML. Use the String set
option when the entire XML file must be read from a previous stage in the
designer canvas. Another stage reads the XML data and passes it on to the XML
Stage.

150 InfoSphere DataStage for Enterprise XML Data Integration

 Use the following steps to configure the String set option in the XML Parser Step:
1. Select the String set option in the XML Source tab. The drop-down list for the
String set option becomes active.
2. Click the String set drop-down list. All input columns to the XML Stage show.
First, top/InputLinks, followed by the name of the input link to the XML
Stage, shows. If the input link is DSLink2, the column name
top/InputLinks/DSLink2/claim_XML shows.
3. Select the claim_XML column from the drop-down list.

After following these steps, you can see that the XML Source tab is configured,
as shown in Figure 7-6 on page 151.

Figure 7-6 The XML Source tab in the XML Parser Step

Chapter 7. Consuming XML documents 151

7.1.2 Reading multiple XML files from a single directory or folder
The healthcare company also stores several of the XML files in a folder on one of
its internal servers. The company does not store these files in the database.
Therefore, these files must be read by the XML Stage directly from the folder.

You can use the File set option in the XML Source tab of the Parser Step to read
multiple files that need to be parsed by the stage. You need to pass the absolute
location of the input XML files to the File set option. The File set option reads the
files from the folder based on the absolute paths of the files.

You can use one of the following approaches to feed the absolute path of the
XML files to the XML Stage:
򐂰 A single flat file can be created to store the absolute path of each input XML
file. This flat file can be read by using the Sequential File stage in the designer
canvas. The input of the sequential file must be given to the XML Stage, as
shown in Figure 7-7. With this approach, you need to make an entry in the flat
file for every new XML file.

Figure 7-7 Sequential_File_Stage reads multiple XML documents job design

򐂰 The External Source stage in the parallel canvas can be used to find the
absolute paths of the XML files. The output of the External Source stage must
be given to the XML Stage, as shown in Figure 7-8. The advantage of using
the External Source stage is that you do not need to enter the paths manually
in a flat file. The stage can be configured to run a query and fetch all XML file
locations at run time from the folder, for example, a query to read the XML
files from the folder C:\files is ls C:\files\*.xml (ls is the command to list
files).

Figure 7-8 External Source stage reads multiple XML documents job design

152 InfoSphere DataStage for Enterprise XML Data Integration

 In both cases, as input is provided to the XML Stage, the Input Step becomes
active and the file_paths column is displayed, as shown in Figure 7-9 on
page 153.

Figure 7-9 File_paths column in the Input Step in the Assembly Editor

Use the following steps to configure the File set option in the XML Parser Step:
1. Select File set option in the XML Source tab. The drop-down list under the
File set option becomes active.
2. Click the drop-down list. The input columns to the XML Stage appear.
All input columns appear beginning with top/InputLinks followed by the
name of the input link to the XML Stage. Therefore, if the input link is DSLink2
and the input column name is file_paths, the column name is
top/InputLinks/DSLink2/file_paths.
3. Select the column file_paths from the drop-down list. The column file_paths
holds the absolute location of the XML files.

Chapter 7. Consuming XML documents 153

 After following these steps, you can see that the XML Source tab is configured,
as shown in Figure 7-10 on page 154.

Figure 7-10 File set option configuration in the XML Parser Step

7.1.3 Reading a single file directly from the XML Stage

The healthcare company can also configure the XML Stage to directly read the
XML file. You can use the Single file option in the XML Source tab of the Parser
step to read an XML file directly from the server. You need to specify the absolute
location of the file and the file name in the Single file option, as shown in
Figure 7-11 on page 155.

154 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-11 Directly read configuration in the XML Parser Step

In this job design, the XML Stage is the source stage in the designer canvas and
no input is provided to it, as shown in Figure 7-12.

Figure 7-12 Job design when the XML file is read from the XML Stage

Chapter 7. Consuming XML documents 155

7.1.4 Specifying the XML schema in the Document Root tab
The XML Stage is a schema-driven stage. Without a valid schema, you cannot
design the assembly. In order for the schema to be available within the assembly,
the schema (xsd files) needs to be imported into the Schema Library Manager.
When you import the schema into the Schema Library Manager, the schema is
converted into a simplified model. This simplified model is then available to the
steps of the assembly.

The fictitious healthcare company based its input files on an XML schema. This
schema needs to be defined in the Document Root tab of the XML Parser Step,
as shown in Figure 7-13. The schema defined in the Document Root tab
describes the common structure of all input XML files to be parsed by the XML
Parser step.

Figure 7-13 Document Root tab in the XML Parser Step

Use the following steps to define the schema in the Document Root tab:
1. Click Browse in the Document Root tab of the Parser Step, and the Select
Document Root window opens. This window shows all the available libraries

156 InfoSphere DataStage for Enterprise XML Data Integration

 in the Schema Library Manager. Each library shows the elements of the
schema that are imported into it.
The healthcare company created two libraries in the Schema Library
Manager, Claim and Customer, and imported the xsd files Claim.xsd and
Customer.xsd into the appropriate libraries. You can see both libraries in the
Select Document Root window, as shown in Figure 7-14.

Figure 7-14 Select Document Root window

2. In the Select Document Root window, select the top-level element on which
the XML file is based. The top-level element in the claim files of the company
is Claim, as shown in Example 7-1. Therefore, you select the item Claim in
the Select Document Root window.

Example 7-1 An example of the Claim XML file

<?xml version="1.0" encoding="UTF-8"?>
<Claim>
<Claim_Information>
<claim_line_number>1</claim_line_number>
<claim_number>A100</claim_number>
<member_id>12Z00</member_id>
<provider_id>00X0</provider_id>
<claim_code>I</claim_code>
</Claim_Information>
</Claim>

3. After you select the element Claim in the Select Document Root window, you
can see the structure of the element on the right side of the window, as shown
in Figure 7-15 on page 158.

Chapter 7. Consuming XML documents 157

 Figure 7-15 The Select Document Root window that shows the schema structure

4. Click OK to close the Select Document Root window. The schema structure is
defined in the Document Root tab, as shown in Figure 7-16 on page 159.

158 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-16 Schema defined in the Document Root tab of the Parser Step

7.1.5 Validating the XML file against the schema

To claim insurance from the healthcare company, the customers need to fill the
claim form through a web application. These claim forms are converted into XML
files. Human errors occur, because the forms are manually completed by the
customers. Therefore, the company validates the claim XML files against the
schema before the data is stored in the database.

Chapter 7. Consuming XML documents 159

 The XML Stage has unique validation capabilities so that you can control the
amount of schema validation when parsing XML documents. The spectrum of
validation ranges from minimal validation to strict schema validation. You control
the trade-off between performance and validation so that you can achieve the
degree of complexity that you want.

The degree of validation can be selected through the Validation tab in the XML
Parser Step, as shown in Figure 7-17. The Validation tab presents a set of
validation rules. Each rule consists of a condition and an action. The condition is
the validation check. If the check fails, the action is performed.

Figure 7-17 The Validation tab in the XML Parser Step

160 InfoSphere DataStage for Enterprise XML Data Integration

The Validation Rules are in two main categories, as shown in Figure 7-17 on
page 160, which are Value Validation and Structure Validation. Value validation
rules check the actual data values in the XML document against the defined
schema types. Structure Validation rules check whether the XML document
conforms to the schema, which is specified in the Document Root tab.

By default, Minimal Validation is selected as the validation mode in the XML

Parser Step. Minimal Validation mode sets all the validation rule actions to
Ignore. The only verifications on the XML document in this mode are that the
top-level element of the XML file matches the top-level element defined in the
Document Root and that the XML document is formed. No other validation is
performed on the XML file.

To ensure that the XML file strictly conforms to the schema, Strict Validation
mode can be selected. In Strict Validation mode, all the validation rule actions are
set to Fatal to ensure that the job cancels on the first occurrence of invalid data.

Each of the validation rules can be configured to perform a specific action. When
you click the action for a particular rule, the drop-down list shows the available
actions that can be selected for that rule, as shown in Figure 7-18 on page 162.

Chapter 7. Consuming XML documents 161

Figure 7-18 The actions available for the validation rules

While entering the claim information in the web application, the customers of the
healthcare company might make errors, such as the following errors:
򐂰 Not provide any data for the Phone Number field
򐂰 Add leading or trailing spaces in the name fields
򐂰 Provide a claim code that is invalid

162 InfoSphere DataStage for Enterprise XML Data Integration

The healthcare company takes the following actions for each of these errors:
򐂰 Provides a default Phone Number
򐂰 Trims all leading and trailing spaces
򐂰 Logs the information when the claim codes are invalid

The default Phone Number can be specified in the XML schema or in the
Administration tab in the Assembly Editor, as shown in Figure 7-19. The Phone
Number field is a string field, and therefore, the default value must be entered in
the text box next to String in the Default Data Type Values tab. To add this default
value to the element during parsing, you must set the action for the validation rule
Use global default values for missing values to Yes.

Figure 7-19 The Default Data Type Values window in the Administration tab

To trim the leading and trailing spaces in each of the fields, you must set the
action for the validation rule Trim Values to Yes.

Chapter 7. Consuming XML documents 163

 To log error messages, the actions Log per Occurrence and Log once per
Document are available. These messages are logged in the DataStage Director
Log. Log per Occurrence logs the message for each data that fails the validation
rule, and Log once per Document logs only the first error message when the
validation rule fails. For example, in the XML document that is shown in
Example 7-2, the claim code fields have invalid data. Log per Occurrence logs
messages for each of the claim codes. Therefore, two error messages are
present in the Director log. Log once per Document logs the message only for
the first invalid claim code, that is AZXBIUTRE.

Example 7-2 An example of the Invalid XML file

<?xml version="1.0" encoding="UTF-8"?>
<Claim>
<Claim_Information>
<claim_line_number>1</claim_line_number>
<claim_number>A100</claim_number>
<member_id>12Z00</member_id>
<provider_id>00X0</provider_id>
<claim_code>AZXBIUTRE</claim_code>
<phone_number></phone_number>
<firstname> James </firstname>
</Claim_Information>

<Claim_Information>
<claim_line_number>2</claim_line_number>
<claim_number>A100</claim_number>
<member_id>12Z00</member_id>
<provider_id>00X0</provider_id>
<claim_code>ZZCVTBJU</claim_code>
<phone_number></phone_number>
<firstname> James </firstname>
</Claim_Information>
</Claim>

The valid claim_code values are specified in the XML schema. To ensure that the
value in the XML file matches the value in the XML schema, the validation rule
Value fails facet constraint is set to Log once per document. The validation rule
Value fails facet constraint ensures that the incoming data value is in the range
specified in the schema.

For the XML document shown in the Example 7-2, if these validation rules are
applied, the following actions occur while parsing:
򐂰 For the Phone Number fields, the default value "1-333-333-0000" is added.
This value needs to be defined in the Administration tab.

164 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 The firstname field is trimmed to output James.
򐂰 An error message is logged in the DataStage Director for the claim_code
field.

The final configuration of the Validation tab can be seen in Figure 7-20.

Figure 7-20 Validation in the Configuration tab in the XML Parser Step

Chapter 7. Consuming XML documents 165

7.1.6 The Reject action
The XML Stage provides an option to reject the invalid claim XML files. To reject
the XML files, the Reject action can be applied to any of the validation rules. The
Reject action turns the status of the validation into regular data on which the
downstream steps or DataStage stages can act.

When the Reject action is selected for any of the validation rules, two new items,
named success and message, are added to the output schema of the XML Parser
step, as shown in Figure 7-21. Success holds the value true if the validation
passes and false if the validation fails. Message holds the error message if the
validation fails. Based on the value of the success item, the XML documents can
be divided into valid and invalid files.

Figure 7-21 The output of the XML Parser step after using the Reject action

For example, assume that the rule Data type contains an illegal value. This rule
checks whether the value specified for each of the elements/attributes in the XML
file is valid and within the range for the specified data type. When the reject
action is selected for this rule and if any of the elements or attributes have invalid
data, the item success holds the value false and message holds the reason for
the failure. Based on the value of success, the files can be filtered into valid and
invalid files. For more detail, see Chapter 9, “Transforming XML documents” on
page 265.

Even if one of the data values in the XML file is invalid, the value of the
successful item is false. Therefore, the entire XML file becomes invalid.

166 InfoSphere DataStage for Enterprise XML Data Integration

 Unlike the other DataStage stages, there might be no reject link from the XML
Stage. All output links from the XML Stage are stream links.

7.1.7 Treating all XML data types as string

The validation rule Data type has an illegal value checks to see whether the
value specified for each of the elements/attributes in the XML file is valid and
within the range for the data type specified. When the action for this rule is set to
Ignore, Log per Occurrence, or Log once per Document, all the elements are
converted to string type. Therefore, the actual XML schema data types are not
validated in the XML Parser step and all elements are treated as string types.

In Figure 7-22 on page 168, we show the actual schema representation in the
Document Root tab of the XML Parser Step. The output of the XML Parser Step
is seen in the Output tab on the rightmost section of the Assembly Editor. The
output of the XML Parser Step shows all elements converted to string elements,
because the action for the validation rule Data type has an illegal value is set to
ignore.

Chapter 7. Consuming XML documents 167

Figure 7-22 Schema representation between Document Root and Output tab

7.1.8 The output of the XML Parser step

To completely configure the XML Parser step, the healthcare company needs to
configure the XML Source, Document Root, and Validation tabs. The output of
the Parser step is seen in the Output tab. The output of the Parser step is
contained within the group XML_Parser:result. You can see the entire XML
schema beneath that group, as shown in Figure 7-23 on page 169.

168 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 7-23 The output of the XML Parser step

When the String set or File set option is selected in the XML Source tab of the
XML Parser Step, the XML_Parser:result is under the input list. As shown in
Figure 7-23, the XML_Parser:result is under the list DSLink2. If the Single file
option is selected, the XML_Parser:result will under the list top.

7.2 The Output step

With the Output step, you can map a hierarchical data structure to a relational
data structure. This step becomes active when output links are present from the
XML Stage. The Output step is used when the XML Stage output needs to be
provided to another stage within DataStage.

The healthcare company wants to parse the XML claim files and store the data in
a database. The Parser step, which we described in 6.1, “The XML Stage” on
page 132, parses the claim files. To write the parsed data into the different
columns of the database, the Output step is required.

Chapter 7. Consuming XML documents 169

 The healthcare company uses a single table into which the data needs to be
written. Because the company uses DB2 as its database, the DB2 Connector
can be used to write the data to the database table. Figure 7-24 shows the
DataStage job design when the XML files are read from the database, parsed,
and written to different tables of the same database.

Figure 7-24 The DataStage job design

When an output link is provided from the XML Stage, the link is represented as a
list within the assembly. The output columns are represented as items in the
assembly. The columns for each of the output links are visible in the Output tab of
the Output Step. As shown in Figure 7-25 on page 171, the output link DSLink4
has five columns in it. These columns can be modified or removed, or new
columns can be added in the Output tab.

170 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-25 The Output tab of the Output Step

To map the output of the Parser step to the columns of the DB2 Connector, the
Mappings tab is used. The Mappings table is a common construct to map one
hierarchical structure to another. The Mappings table is used as the last step in
adapting the input to the output structure. You cannot perform a join, union, or
any other set-operation and map to a target list. If a set-operation is needed, it
needs to be performed in a Transformation step before the Output step. As
shown in Figure 7-26 on page 172, the Mappings table contains three columns:
Source, Result, and Target. The Target column displays the output link and the
columns under it. The Source column displays the columns that need to be
mapped to the target items. The Result column displays the status of the current
mapping.

Chapter 7. Consuming XML documents 171

Figure 7-26 The Mappings tab of the Output step

The top list is always the first list in the Mappings table and is always mapped to
the source item top. This mapping cannot be configured. Top contains the group
OutputLinks. The entire output structure is contained in this group.

To map a hierarchical structure to a relational structure, use the following

mapping rules:
򐂰 You can only map a list in the Target column to a list in the Source column.
For example, in Figure 7-26, only a source list can be mapped to DSLink4.
Because all lists are represented with a blue icon, you must map an item with
a blue icon to another item with a blue icon.
򐂰 A group is never mapped. Groups are only structures used to create, and they
do not need to be mapped. In the Output step, only a single group exists in
the Target column, for example, OutputLinks. No mapping occurs in the
Source column for this group.

172 InfoSphere DataStage for Enterprise XML Data Integration

򐂰 To map a source item to a target column, the parent list of the item needs to
be mapped first. Therefore, in Figure 7-26 on page 172, to map a source item
to the Column claim_line, the list DSLink4 needs to be mapped first.

Following these rules, the output of the Parser step must be mapped to the target
columns in the Output step, as shown in Figure 7-27.

Figure 7-27 Mapping Parser step output and the target columns in the Output step

Use the following steps to complete the Mappings table:

1. Double-click the Source column for the target list DSLink4 to see a
drop-down list. This drop-down list is called the suggestion list. The
suggestion list shows all the available source lists that can be mapped to the
target list. Both source lists, User_information and Claim_Information, are
available in the suggestion list, as shown in Figure 7-28 on page 174. The
absolute paths can be seen in the suggestion list, for example,
top/InputLinks/DSLink2/XML_Parser:result/Claim/User_Information.

Chapter 7. Consuming XML documents 173

Figure 7-28 The suggestion list for the output link DSLink4

2. Because the source columns that need to be mapped to the target columns
are under the source list Claim_Information, this column is selected from the
suggestion list. When you select Claim_Information, the mapping occurs for
this row and a white check mark in a green circle shows between the source
and target lists, as shown in Figure 7-29 on page 175.

174 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-29 The Mappings table for the Source and Target lists

3. After mapping the target list DSLink4, the columns under it can be mapped.
To map each of the target columns, select the corresponding row in the
Source column. From the suggestion list, select the appropriate item to be
mapped. The item recommendations in the suggestion list are based on the
name and data type of the Target column.

After mapping all the columns, the Mappings table is complete. No errors are
seen in the Output Step, as shown in Figure 7-30 on page 176.

Chapter 7. Consuming XML documents 175

Figure 7-30 The completed Mappings table

7.2.1 Automatically mapping the columns

One of the company output tables has over a hundred columns. It can be tedious
to map each column in the target to the corresponding column in the source. The
Output step provides an easy mechanism to map multiple columns
simultaneously, and that is the Auto Map feature. The following steps describe
how to use the Auto map feature:
1. Map the source list Claim_Information to the target list DSLink4.
2. Select the row corresponding to the target list DSLink4.
3. Click Auto Map. All the columns under the list Claim_Information are
automatically mapped.

The mapping occurs based on the similarity between the name and data type of
the source and target columns. The automatic mapping suggestion might not
always be correct. The Auto Map feature also cannot map columns of different
data types. Therefore, if the Target column is an integer and all source columns

176 InfoSphere DataStage for Enterprise XML Data Integration

 are of type string, by using the Auto Map feature, no mapping occurs for the
integer column.

7.2.2 Propagating columns from the source to the target

Sometimes the output columns are not defined in the Output stage as with the
database. The developers at the company need to create and map the columns.
The Output step provides the Propagate feature to facilitate this process. The
Propagate option creates the target columns and also automatically maps these
columns to the appropriate columns in the source structure.

Follow these steps to use the Propagate feature:

1. Map the source list to the target list. For example, DSLink4 is mapped to
Claim_Information in Figure 7-31. No target columns show under the target
list.

Figure 7-31 The Mappings tab

2. Select the row corresponding to the target list DSLink4, and click Propagate.

Chapter 7. Consuming XML documents 177

 3. All the columns under the list Claim_Information are propagated to the target,
as shown in Figure 7-32. The names of the columns created in the target list
are the same names of the columns in the source list.

Figure 7-32 Columns created with the Propagate option

7.2.3 Mapping constants to a target column

The company uses a table in which the ProviderID must always be set to '00X0'
irrespective of the data that comes from the XML files. Therefore, a constant
value needs to be assigned to the output column ProviderID. Use the Output
Step Constant option. To map the target item ProviderID to the constant value
'00X0', use the following steps:
1. Ensure that the parent list of the target item is mapped. Therefore, in
Figure 7-33 on page 179, to assign a constant value to the item ProviderID,
DSLink4 needs to be mapped. So, you can map Claim_Information to the list
DSLink4.

178 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-33 The Mappings table in the Output Step

2. Select the row in the source corresponding to the target item ProviderID. The
suggestion list is shown in Figure 7-34 on page 180.

Chapter 7. Consuming XML documents 179

Figure 7-34 The suggestion list

3. The last option in the suggestion list is Constant. Select the Constant option
to open the Constant window. In the Constant window, as shown in
Figure 7-35, you can enter the value that needs to be assigned to the target
item. To assign the value '00X0', enter '00X0' in the Constant window and
click OK.

Figure 7-35 The Constant window

180 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 7-36 shows a constant value assigned to the Target column ProviderID.

Figure 7-36 Mappings table after assigning a constant value to a column

7.2.4 Selecting mapping candidates by using the More option

When many columns are present in the output of the XML Parser Step, the
suggestion list cannot list all these columns. By default, only five source items
can be seen in the suggestion list. If the required column is not available in the
suggestion list, use the More option. The More option shows the entire source
tree structure from which the Source column to be mapped can be selected.
Follow these steps to use the More option:
1. Select the row in the Source column that needs to be mapped to the Target
column.
2. The More option is available at the end of the suggestion list, as shown in
Figure 7-37 on page 182. Select More.

Chapter 7. Consuming XML documents 181

Figure 7-37 The More option in the suggestion list

3. The Target window opens, as shown in Figure 7-38 on page 183. In this
window, you can see the entire schema tree and select the item to map.

182 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 7-38 The Target window

4. Select the item to map and click OK.

The mapping is complete.

7.3 Mapping suggestions

In the Output step, particular mappings must be performed in certain ways. Two
scenarios are commonly used conditions. We describe these conditions in the
following sections.

7.3.1 Mapping child and parent items to same output list

In the Output step, to map a source item under a list, the list needs to be mapped
to the output link. For example, in Figure 7-39 on page 184, to map the source
item claim_line_number to the Target column claim_line, the list
Claim_information needs to be mapped to the output link DSLink4. If
User_Information is mapped to the output link, only the items under this link can
be mapped to the target items and claim_line_number cannot be mapped to the
target item.

Chapter 7. Consuming XML documents 183

 Figure 7-39 The source and target structures

In certain instances, you might want to map the parent and child fields to
columns under the same output link in the Output step. This way, you relate the
parent and child columns in the output file or database. For example, in
Figure 7-40, the source items claim_type and claim_line_number can be
mapped to the columns under the same output link. These source items are not
under the same list. For example, the claim_type item is the child of
Claim_details, and Claim_details is the parent of Claim_Information.

Figure 7-40 The source and target structures

184 InfoSphere DataStage for Enterprise XML Data Integration

 To map two columns under different lists to the same output link, the innermost
list is mapped to the output link. Therefore, the list Claim_Information is mapped
to the output link DSLink4. Now, you can map both items, claim_line_number and
claim_type, to the appropriate output columns. If the list Claim_details is mapped
to the output link, only the item claim_type can be mapped to the output column.
You cannot map claim_line_number.

It is important that the lists Claim_details and Claim_Information share a

parent-child relationship. If these lists are in the same level, such as the lists
Claim_Information and User_Information in Figure 7-39 on page 184, the
columns under these lists cannot be mapped to a single output link.

7.3.2 Mapping when no lists are available in the Parser step

Certain schemas do not include lists. That is, no element occurs more than one
time in the schema. For example, the schema that is shown in Figure 7-41 does
not show any multiple occurrences of the same element. You want to map the
items of this XML schema to the columns of a database by using the Output step.

Figure 7-41 The source and target structures

To map the source columns to the target columns, first the source list and target
list need to be mapped. But as you can see in Figure 7-41, no lists show in the
output of the XML Parser step. To map them, use the following options.

Group to List projection

The Group to List projection is available in the Output tab of each step as part of
the other projections, such as Drop or Rename. The Group To List projection can
apply to a group node (an element with a brown icon) only. For more detail, see
Chapter 5, “Schema Library Manager” on page 113. This option converts the
group into a list of one.

Chapter 7. Consuming XML documents 185

 You can use the Group to List projection to create a list in the output schema of
the XML Parser Step. Follow these steps:
1. Right-click the group XML_Parser:result in the Output tab of the XML Parser
Step. The Group to List projection is visible, as shown in Figure 7-42.

Figure 7-42 The Group to List projection

2. Select the Group To List projection. The group changes into a list, as shown
in Figure 7-43 on page 187.

186 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 7-43 The XML_Parser:result as a list

3. In the Output step, XML_Parser:result can be mapped to the output list

DSLink4, as shown in Figure 7-44.

Figure 7-44 The complete mapping between the source and target structures

Mapping top to the output list

The entire input and output tree in the assembly is present under the list top. The
list top is the parent list in the assembly. The result of the XML Parser step is also
present under the list top.

When the Single File option is selected in the XML Source tab of the XML Parser
Step, the output of the XML Parser Step is directly under the list top. Therefore,
in the Output Step, the list top can be mapped to the output link to map the items
under the XML Parser Step, as shown in Figure 7-45 on page 188. The warnings
in the mapping are the result of the difference in data types between the Source
column and the Target column.

Chapter 7. Consuming XML documents 187

Figure 7-45 The Mappings table

7.4 Parsing large schemas

Many schema types are available. Parsing the types of schemas works the same
way as the XML Parser step. However, you must change the job design for large
schemas.

The XML Stage is a schema-driven stage. To parse any XML document, the
schema needs to be defined in the Document Root tab of the XML Parser Step.
This XML schema is converted into an internal model by the XML Stage. When
large schemas are used, a large amount of memory is required to store the
schema in the internal XML Stage model, because every aspect of the schema
needs to be saved. Therefore, for large schemas, the representation of the XML
schema in the assembly is changed.

For large schemas, several elements are automatically chunked. Chunking refers
to converting all child items under an element to a single item.

The leftmost diagram in Figure 7-46 on page 189 shows the actual schema
representation. When this schema is part of another large schema, the element
Claim_Information might be automatically chunked. As shown in Figure 7-46 on

188 InfoSphere DataStage for Enterprise XML Data Integration

page 189, the auto-chunked element is represented with the symbol <>. All child
items of the element Claim_Information are contained in this element.

Figure 7-46 Schema structure when a node is auto-chunked

When an element is automatically chunked, its child items are not available for
mapping in the Output step. For example, you cannot map the element
claim_line_number to the Target column in the Output step, because the Output
step sees only the single element e2res:text() under Claim_Information.

To parse the automatically chunked elements, another XML Parser step is

required in the assembly. The second XML Parser step is used to parse the
auto-chunked element so that the child elements can be available for mapping in
the Output step.

Follow these steps to parse the auto-chunked element:

1. Add a second XML Parser step to the assembly.
2. In the second Parser step, use the String Set option to read the auto-chunked
element. Select the String Set option in the XML Source tab. From the
drop-down list, select e2res:text(), which is the auto-chunked node, as shown
in Figure 7-46. The String Set option reads the XML data from the first XML
Parser step.
3. As soon as the String Set option is configured, the Document Root is
automatically set by the XML Stage. Therefore, you do not need to configure
the Document Root tab for the second XML Parser step. The Document Root
is the same structure as the auto-chunked node, as shown in Figure 7-47 on
page 190.

Chapter 7. Consuming XML documents 189

Figure 7-47 Document Root tab that shows the structure of the auto-chunked element

After configuring the second XML Parser step, the child items of the
auto-chunked element are available for mapping in the Output step.

7.4.1 Output schema location of the second XML Parser step

The output of the second XML Parser step is under the parent list of the
auto-chunked element, as shown in the following examples:
򐂰 Look at the auto-chunked element under the list Claim_Information in
Figure 7-48 on page 191. The output of the second XML Parser step shows
under the list Claim_Information, because Claim_Information is the parent list
of the auto-chunked node.

190 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 7-48 Output schema of the second XML Parser step

򐂰 Assume that the auto-chunked element, book, is under the group library. The
parent list of the element book is top. Therefore, the output of the second
XML Parser step is placed under the list top, as shown in Figure 7-49 on
page 192.

Chapter 7. Consuming XML documents 191

 Figure 7-49 Output schema of the second XML Parser step

򐂰 Assume that the auto-chunked element phone_number is under a group

address. This group is under the Employee list. Because the parent list of the
element phone_number is Employee, the output of the second XML Parser
step is under the Employee list, as shown in Figure 7-50.

Figure 7-50 Output schema of the second XML Parser step

192 InfoSphere DataStage for Enterprise XML Data Integration

7.4.2 Configuring the auto-chunked elements in the assembly
While configuring the Document Root in the XML Parser step, select the top-level
element from the Schema Library Manager. The top-level element is in the form
of a schema tree, containing schema elements in a hierarchical structure. For
every top-level element, the XML Stage maintains a count of the number of
schema elements imported into the assembly.

Two fields are defined in the Administration tab, as shown in Figure 7-51. You can
use these fields, Schema Trimming starts at and Maximal Schema Tree size, to
specify thresholds on the expected size of schema trees that need to be
imported. When the number of imported elements reaches these thresholds,
XML Stage starts auto-chunking the schema elements. These thresholds are
soft limits.

Figure 7-51 The Administration Configuration tab

The XML Stage starts to chunk the top-level XML elements when the tree size
reaches the value for the Schema Trimming starts at. When the tree size
reaches Maximal Schema Tree size, the XML Stage auto-chunks all the qualified

Chapter 7. Consuming XML documents 193

 XML elements (not only the top-level XML elements, but also nested XML
elements).

If you do not use a large schema inside the assembly, there is no user behavior
change. The field Schema trimming starts at defaults to 500, and the field
Maximal Schema Tree size defaults to 2000. For schemas whose size is smaller
than 500, XML Stage does not auto-chunk any schema elements. When you
need to import a large schema into the assembly, configure the two fields in the
Administration tab and use multiple parsers to extract the auto-chunked
elements.

Because the values are defined inside the Assembly Editor, they apply to the
particular assembly in use. When the schema is defined in the Document Root
tab of the XML Parser step, the representation of the schema depends on the
values for the Schema Tree Configuration fields at that instance. After defining
the schema in the Document Root of the XML Parser step, the values can be
changed. The changed values do not affect the schema representation in the
XML Parser steps in which the Document Root is already defined. Therefore, in a
single assembly, each XML Parser step can be configured to use different
configuration values.

7.4.3 Designing the assembly

When the auto-chunked element is selected in the second XML Parser step, the
Document Root is automatically defined. The Document Root is automatically
defined if the first and second XML Parser steps are defined one after another
without saving the assembly in between. But, in case, the Assembly Editor is
closed after defining the first XML Parser step, the Document Root is not
automatically defined in the second XML Parser step when it is added to the
assembly on reopening the Assembly Editor.

For example, suppose that you configure the first XML Parser step, save the
Assembly Editor by closing it, and then reopen it. When you add the second XML
Parser step to parse the auto-chunked element, the Document Root is not
automatically defined. You must go to the first parser and set the Document Root
again. You do not need to delete or reconfigure the first parser. You merely need
to reset the Document Root.

This step also applies to a scenario with several parsers in the assembly. The
first XML Parser is used to import the whole schema. The rest of the Parsers are
used to parse several auto-chunked elements from the output of the first parser.
If you reopen the job to add another parser to parse another auto-chunked
schema element, you need to reset the Document Root in the first parser.

194 InfoSphere DataStage for Enterprise XML Data Integration

7.5 Parsing only a section of the schema
Occasionally, only a part of an XML document must be written into the output,
along with a particular XML element. The healthcare company stores most of the
XML files in the database. However, the company wants to extract an element
from the XML file and then store the XML file and the extracted element in two
columns of another database table.

Use the XML Parser step to parse the XML file, extract a single element from it,
and write the XML file along with the column back into the database.

Use the chunk operation in the XML Parser step to perform these operations.
This operation is manual as opposed to the automatic chunk described in 7.4,
“Parsing large schemas” on page 188.

The chunk feature is available in the Document Root tab of the XML Parser step.
You can select any element in the Document Root window and right-click that
element to activate the chunk option, as shown in Figure 7-52 on page 196.

Chapter 7. Consuming XML documents 195

Figure 7-52 The Chunk option in the Document Root tab of the XML Parser Step

Chunking refers to converting all child items under an element to a single item. In
Figure 7-53 on page 197, we show the actual schema representation. When the
element Claim is manually chunked, it is represented with the symbol <>. All the
child items of the element Claim are contained in this single element, as shown in
Figure 7-53 on page 197.

196 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-53 Manual chunk representation

To extract the item claim_number, the chunk Claim needs to be passed through a
second XML Parser step. Use the following steps to parse the element Claim:
1. Add a second XML Parser step to the assembly.
2. In the second XML Parser step, use the String Set option to read the chunked
element. Select the String Set option in the XML Source tab. From the
drop-down list, select Claim. The String Set option reads the XML data from
the first XML Parser step.
3. As soon as the String Set option is configured, the Document Root is
automatically set by the XML Stage. Therefore, you do not need to configure
the Document Root tab for the second XML Parser step.

After configuring the second Parser step, the chunked element and its child items
are available for mapping in the Output step, as shown in Figure 7-54. You can
map the items Claim and claim_number to two columns in the Output step.

Figure 7-54 Output available to the Output step

Chapter 7. Consuming XML documents 197

 The output of the second XML Parser step is under the parent list of the chunked
element. See 7.4.1, “Output schema location of the second XML Parser step” on
page 190 for a detailed description of the location of the output schema for
various types of schemas.

For auto-chunked elements, to get the Document Root automatically defined in

the second XML Parser step, the first and second Parser steps need to be
defined at the same time in the assembly. However, the first and second Parser
steps do not need to be defined at the same time in the assembly for manually
chunked elements. Even when you save and reopen the assembly, the
Document Root gets automatically defined in the second XML Parser step.

7.6 Incorporating an XSLT style sheet during parsing

The Assembly Editor does not require you to create an Extensible Stylesheet
Language Transformation (XSLT) style sheet, but instead, it provides a more
intuitive way to describe the transformation. However, there are circumstances
when you need to first employ an XSLT style sheet on the data and then parse it
and map it in the assembly. We do not suggest this approach for large
documents. The current XSLT technology requires the document to fit in memory.
It therefore cancels the advantage of the assembly that can parse any document
size.

However, if necessary, the assembly can first perform the XSLT transformation,
and it reads the result of the XSLT as though it is the content of the original file.
Therefore, the Document Root element reflects the result of the XSLT and not the
original XML input document, as shown in Figure 7-55.

XSLT
Input XML File XML Parser Step
Stylesheet
XML Stage

Figure 7-55 Processing XML documents that incorporate XSLT style sheets

You must paste the XSLT style sheet in the Enable Filtering section of the XML
Source window in the XML Parser step. To enable the Enable Filtering option,
first the XML Source needs to be defined, and then, you must select Enable
Filtering. A default XSLT style sheet shows in the area under the Enable
Filtering option, as shown in Figure 7-56 on page 199. Without selecting an XML
Source, the Enable Filtering option is not enabled. You can replace the default
style sheet with a style sheet of your choice.

198 InfoSphere DataStage for Enterprise XML Data Integration

Figure 7-56 Enable Filtering option in the XML Source tab of the XML Parser Step

The healthcare company uses several XML files that do not conform to the XML
schema. That is, the top level of the XML file and the schema do not match. But
the company still wants to parse these files by using the XML Stage. To ensure
that the XML documents conform to the XML schema, the company wants to
apply an XSLT style sheet to the XML documents. The XSLT must be pasted in

Chapter 7. Consuming XML documents 199

 the Enable Filtering area in the XML Source tab of the XML Parser Step. And, the
output of the XSLT must conform to the XML schema defined in the Document
Root of the Parser Step, as shown in Figure 7-55 on page 198.

200 InfoSphere DataStage for Enterprise XML Data Integration

 8

Chapter 8. Creating and composing

XML documents
Our fictitious example bank is in New York. The bank was established in the year
1990 and provides services for many global IT companies. The bank has over a
million account holders. The bank stores all customer information in a database.
The bank sends a report to its customers every quarter. This report summarizes
the activity of the customers in that quarter. The application that creates these
reports needs all the information in an XML document.

Customer Application to
Information in Creating XML files from create reports
the database the data in the database
Customers

Figure 8-1 The bank configuration

© Copyright IBM Corp. 2012. All rights reserved. 201

8.1 The XML Composer step
The bank stores all its customer information in a database. By using the data in
the database tables, an XML file needs to be created to be input to the
application that creates the reports. The bank already uses IBM Information
Server for managing its business requirements and analyzing its business terms.
Therefore, the bank wants to use InfoSphere DataStage to create the XML files.

8.1.1 Writing XML files

To create the XML files, the XML Output stage or the XML Stage can be used in
the DataStage Designer. Because it is easier to create large XML documents
with multiple levels of hierarchy by using the XML Stage, XML Stage is the
preferred choice for creating the XML files. In the XML Stage, you use the
Composer step to create the XML files. To add the Composer step to the
Assembly Editor, follow these steps:
1. Open Assembly Editor and click Palette in the Assembly Outline bar. The
steps show in the Palette, as shown in Figure 8-2 on page 203.

202 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-2 The Palette in the Assembly Editor

2. Double-click XML_Composer in the Palette. The XML_Composer Step is

added between the Input Step and the Output Step, as shown in Figure 8-3
on page 204.

Chapter 8. Creating and composing XML documents 203

Figure 8-3 The XML_Composer Step added to the Assembly Editor

The bank uses IBM DB2 as the database storage. The customer information is
stored in multiple columns in a single database table. Therefore, in the
DataStage job, the source of input data is DB2. In the DataStage Designer
canvas, the DB2_Connector_Stage can be used to read the relational columns
from the database. The output of the DB2_Connector_Stage must be fed to the
XML Stage. The overall job design is shown in Figure 8-4.

Figure 8-4 The DataStage job to read data from DB2 and create an XML file

204 InfoSphere DataStage for Enterprise XML Data Integration

 The Input Step in the Assembly Editor describes the metadata of a relational
structure. The step becomes active when an input link is provided to the XML
Stage. You can see the columns that are added to the preceding stage of the
XML Stage. The columns can be modified or removed, or new columns can be
added. When the input is provided from the DB2 Connector to the XML Stage,
the Input Step becomes active and displays the DB2 columns, as shown in
Figure 8-5.

Figure 8-5 The Input Step in the Assembly Editor

The bank wants to store the output XML files in two kinds of columns in the
database: in a pureXML column and in a large object column. The Composer
step writes the XML files in these locations. In the XML Composer step, the
method of writing the XML file must be specified in the XML Target tab. A single
Composer step cannot write the XML data to multiple destinations at a time.

Writing XML files to a pureXML column in the database

The bank wants to store the XML files in a single DB2 pureXML column in the
database. In a pureXML column, the XML files can be stored in their native

Chapter 8. Creating and composing XML documents 205

 hierarchical format without the need to shred the data into relational format.
Applications can readily access the XML data in this column.

In the DataStage Designer canvas, the DB2 Connector stage can be used to
write the XML files into the database table. The XML files are stored in a single
column xml_data in the database. The DB2 Connector can access the correct
database table to write into this particular column. Because the XML Stage
needs to pass the created XML files to the DB2 Connector stage, the output of
the XML Stage must be given to the DB2 Connector stage, as shown in
Figure 8-6.

Figure 8-6 DataStage job design: XML file written into pureXML column of the database

To pass the XML files to a downstream stage within DataStage, you must use the
Pass as String option in the XML Target tab of the Composer step. When the
Pass as String option is used, the Composer step creates the XML files, and
another stage within DataStage actually writes these files to the target.

Select Pass as String in the XML Target tab, as shown in Figure 8-7 on
page 207. The XML Target tab is then configured to pass the created XML files to
a downstream stage.

206 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-7 The XML Target tab in the Composer Step

Writing XML files to a large object column in the database

The XML files can be stored as large object (LOB) columns in the DB2 database.
We use LOB to refer to LOB, BLOB, CLOB, or DBCLOB data types in the
database. See 10.6.2, “Passing XML by reference as a large object column” on
page 321, where LOB is described in detail.

To pass the XML files to a LOB column in the DB2 Connector stage, the Pass as
Large Object option must be used in the XML Target tab of the Composer Step.
The Pass as Large Object option is used to pass the XML files to a LOB-aware
stage. Therefore, the file is created in the Composer step and passed to a
LOB-aware stage in DataStage, which actually writes the file to the target. The
job design is shown in Figure 8-8.

Figure 8-8 DataStage job design: XML file is written to LOB column of the database

Chapter 8. Creating and composing XML documents 207

 Select Pass as Large Object in the XML Target tab, as shown in Figure 8-9. The
XML Target tab is then configured to pass the created XML files to a downstream
LOB-aware stage.

Figure 8-9 The XML Target tab LOB option in the Composer Step

Writing the XML file from the Composer step

The bank can also configure the XML Composer step to directly write the XML
file to the server. The Write to File option can be used to write the XML file to the
server. Select Write to File. You need to specify the absolute location where the
XML files are created, as well as the file name. In the Output Directory text box,
enter the absolute location where the file is to be created. Enter the XML file
name in the Filename Prefix text box. For example, as shown in Figure 8-10 on
page 209, to create the XML file file1.xml in the C:\files directory, you must
enter C:\files for Output Directory. Enter file1 for Filename Prefix. When you
enter the file name, do not type the extension .xml. The extension is added by the
Composer step when you create the file.

208 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-10 The XML Target tab Write to file option in the Composer Step

In this job design, the XML Stage is the target stage in the designer canvas. No
output is provided from the XML Stage, as shown in Figure 8-11.

Figure 8-11 DataStage Job Design: XML file is written from the XML Stage

8.1.2 Specifying the XML schema in the Document Root tab

The XML Stage is a schema driven stage. Without a valid schema, you cannot
design the assembly. For the schema to be available in the assembly, the
schema (xsd files) must be imported into the Schema Library Manager. When
you import the schema into the Schema Library Manager, the schema is
converted into a simplified model. This simplified model is then available to the
steps of the assembly.

Chapter 8. Creating and composing XML documents 209

 The bank must create XML files from the data in the database. To create XML
files, the XML schema must be defined in the Document Root tab of the
Composer Step, as shown in Figure 8-12. The schema defined in the Document
Root tab describes the common structure of all output XML files to be created by
the Composer Step.

Figure 8-12 The Document Root tab of the Composer Step

To define the schema in the Document Root tab, follow these steps:
1. Click Browse. The Select Document Root window opens. This window shows
all the available libraries in the Schema Library Manager. Each library shows
the elements of the schema that are imported into it.

The bank created the library Account_Info in the Schema Library Manager and
imported the xsd file Account.xsd into it. When you click Browse in the
Document Root tab of the Composer Step, you can see the library in the Select
Document Root window, as shown in Figure 8-13 on page 211.

210 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-13 The Select Document Root window

2. In the Select Document Root window, select the top-level element of the XML
file. The output XML files of the bank must have Account as their top-level
element, as shown in Example 8-1. Therefore, you must select the item
Account in the Select Document Root window.

Example 8-1 Expected output XML file

?xml version="1.0" encoding="UTF-8"?>
<Account>
<Customer>
<Account_Number>C198765487</Account_Number>
<Account_Holder_Name>John</Account_Holder_Name>
<Account_Holder_DOB>1980-07-02</Account_Holder_DOB>
<Account_Holder_address>San Jose,CA</Account_Holder_address>
<Account_Holder_phoneNumber>123-456-654-4567</Account_Holder_p
honeNumber>
<Account_Type>Savings</Account_Type>
</Customer>
</Account>

3. When you select the element Account in the Select Document Root window,
you can see the structure of the element on the right half of the window, as
shown in Figure 8-14 on page 212.

Chapter 8. Creating and composing XML documents 211

 Figure 8-14 The schema structure in the Select Document Root window

4. Click OK to close the Select Document Root window. The schema structure is
defined in the Document Root tab, as shown in Figure 8-15 on page 213.

212 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-15 The Document Root tab in the Composer Step

8.1.3 Creating a valid XML file

The application that creates the report for the bank requires data in an XML file.
The XML file needs to be valid, or the application cannot create the report.
Therefore, the bank wants to validate the XML file before the file is input to the
application that creates the reports.

The XML Stage offers unique validation capabilities for you to control the amount
of schema validation when composing XML documents. When the XML
Composer step produces the XML document, it always creates a well-formed
document. It also validates the incoming data from the previous steps against the
document root element structure. The spectrum of validation that ranges from
minimal validation to strict schema validation allows control of the trade-off

Chapter 8. Creating and composing XML documents 213

 between performance and validation and allows the degree of complexity that
you want.

The degree of validation can be selected through the Validation tab in the XML
Composer Step, as shown in Figure 8-16. The Validation tab presents a set of
validation rules. Each rule consists of a condition and an action. The condition is
the validation check, and when the check fails, the action is performed.

Figure 8-16 The Validation tab in the Composer Step

The Validation Rules are categorized into two categories, as shown in

Figure 8-16: Value Validation and Structure Validation. Value validation rules
are rules that check the actual data values in the XML document against the
defined schema types. Structure Validation rules are rules that check whether
the XML document conforms to the schema that was specified in the Document
Root tab.

214 InfoSphere DataStage for Enterprise XML Data Integration

 Not all rules are available for XML Composer step. Certain rules are only
applicable for the XML Parser step, because the XML Composer step always
adheres to these rules.

By default, Strict validation is selected as the validation mode in the Composer

step. In Strict validation, all the validation rule actions are set to Fatal to ensure
that the job stops on the first occurrence of invalid data.

Minimal validation mode sets all the validation rule actions to Ignore. When
Minimal validation is selected in the Composer step, no validation occurs. The
Composer step, by default, always produces a well-formed XML document. Even
when Minimal Validation is selected, the Composer step creates a well-formed
XML document.

Each of the validation rules can be configured to perform a specific action. When
you click the action for a particular rule, the drop-down list shows the available
actions that can be selected for that rule, as shown in Figure 8-17.

Figure 8-17 The actions that are available for the validation rules

Chapter 8. Creating and composing XML documents 215

 The bank uses the data from the database while creating the XML files. Because
the data in the database is stored in varchar columns, some of the data can be
invalid. You might see the following issues in the incoming data:
򐂰 Incorrect format for the date of birth (DOB) field. The correct date format in
XML is yyyy-mm-dd.
򐂰 The Account Type field might not be the correct type.

The bank wants to take the following action for each of these errors:
򐂰 Log an error when the date formats are incorrect.
򐂰 Log an error when the Account Type field is an incorrect value.

To log error messages, the actions Log per Occurrence and Log once per
Document are available. These messages are logged in the DataStage Director
Log. Log per Occurrence logs the message for data that fails the validation rule.
Log once per Document logs the first error message only when the validation
rule fails. For example, in the XML document shown in Example 8-2, the Account
Type fields contain invalid data. Log per Occurrence logs messages for each of
the account types, and therefore, two error messages show in the Director log.
Log once per Document logs the message only for the first invalid account type,
for example, Private.

Example 8-2 Output XML file created by the Composer step

<?xml version="1.0" encoding="UTF-8"?>
<Account>
<Customer>
<Account_Number>C198765487</Account_Number>
<Account_Holder_Name>John Doe1</Account_Holder_Name>
<Account_Holder_DOB>02-07-1980</Account_Holder_DOB>
<Account_Holder_address>San Jose,CA</Account_Holder_address>
<Account_Holder_phoneNumber>123-456-654-4567</Account_Holder_phon
eNumber>
<Account_Type>Private</Account_Type>
</Customer>
<Customer>
<Account_Number>C198765444</Account_Number>
<Account_Holder_Name>Jane Smith</Account_Holder_Name>
<Account_Holder_DOB>1975-01-01</Account_Holder_DOB>
<Account_Holder_address>San Jose,CA</Account_Holder_address>
<Account_Holder_phoneNumber>123-789-987-2457</Account_Holder_phon
eNumber>
<Account_Type>non-savings</Account_Type>
</Customer>
</Account>

216 InfoSphere DataStage for Enterprise XML Data Integration

The validation rule Data type has an illegal value checks whether the value
specified for each of the elements/attributes in the XML file is valid and within the
range for the specified data type. This rule checks the format of the incoming
data for the date of birth field. Set the action for this rule to Log per Occurrence to
log an error message for every invalid date format.

The valid Account Type values are specified in the XML schema. The data that
comes from the database must match the value in the XML schema. If the data
does not match, an error message needs to be logged. To log an error message
for every invalid Account Type, the validation rule Value fails facet constraint is
set to Log per Occurrence.

For the XML document that is shown in Example 8-2 on page 216, if these
validation rules are applied, the following results occur while parsing:
򐂰 An error message is logged for the date of birth field, because the value
02-07-1980 is not in the correct format.
򐂰 Two error messages are logged for the Account Type field for the values
Private and non-savings.

You can see the final configuration of the Validation tab in Figure 8-18 on
page 218.

Chapter 8. Creating and composing XML documents 217

Figure 8-18 Configuration of the Validation tab in the XML Composer Step

Treating all XML data types as string

The validation rule Data type has an illegal value checks whether the value
specified for each of the elements/attributes in the XML file is valid and within the
range for the specified data type. When the action for this rule is set to Ignore,
Log per Occurrence, or Log once per Document, then all the elements are
converted to string type. Therefore, the actual XML schema data types are not
validated in the Composer step and all elements are treated as a string. For
example, if an element is of integer type, and Data type has an illegal value is set
to Ignore, the element is treated as a string type.

8.1.4 The Mappings tab

To map the output of the Input step to the Composer step, the Mappings tab is
used. The Mappings table is a common construct to map one hierarchical
structure to another. The Mappings table is used as the last step in adapting the

218 InfoSphere DataStage for Enterprise XML Data Integration

 input to the output structure. You cannot perform a join, union, or any other set
operation and map to a target list. If a set operation is needed, it must be
performed in a Transformation step before the Composer step.

The Mappings table, which is shown in Figure 8-19, contains three columns:
Source, Result, and Target. The Target column displays the items of the target
structure. The target structure is the XML schema that is defined in the
Document Root tab of the Composer Step. The Source column displays the
columns that need to be mapped to the target items. The Result column displays
the status of the current mapping.

Figure 8-19 The Mappings tab in the Composer Step

The document_collection list is the first target list in the Mappings table. Based
on the mapping to this list, one or multiple XML documents are created (see 8.3,
“Multiple XML documents: Single Composer step” on page 253).

Chapter 8. Creating and composing XML documents 219

 To create only one output XML document, the top list in the Source column
needs to be mapped to document_collection in the Target column. The entire
output schema structure is displayed under the document_collection list, as
shown in Figure 8-20. This structure is similar to the structure that is defined in
the Document Root tab of the Composer Step.

In order to map the items from the Input step to the Composer step, follow these
mapping rules:
򐂰 A list in the Target column can be mapped to a list in the source column only.
For example, in Figure 8-20, only a source list can be mapped to Customer.
All lists are represented with blue icons. An item with a blue icon must be
mapped to another item with a blue icon.
򐂰 A group is never mapped. Groups create structures and do not need to be
mapped. As shown in Figure 8-20, the group Account does not need to be
mapped.
򐂰 To map a source item to a target column, first the parent list of the item needs
to be mapped. Therefore, in Figure 8-20, to map a source item to
Account_Number, first the Customer list needs to be mapped.

Following these rules, the output of the Input step must be mapped to the target
columns in the Composer step, as shown in Figure 8-20.

Figure 8-20 Mapping the Input Step output and the Composer Step target structure

To complete the Mappings table, follow these steps:

1. Double-click the Source column for the target list Customer to see a
drop-down list. This drop-down list is called a suggestion list. The suggestion
list shows all the available source lists that can be mapped to the target list.
Only one source list, DSLink2, exists. Therefore, it is the only available source

220 InfoSphere DataStage for Enterprise XML Data Integration

 list in the suggestion list that is shown in Figure 8-21. The absolute path
shows in the suggestion list, that is, top/InputLinks/DSLink2.

Figure 8-21 The suggestion list for the target list Customer

2. Select the source list DSLink2 from the suggestion list. The mapping occurs
and a white check mark in a green circle shows in the Result column, as
shown in Figure 8-22 on page 222.

Chapter 8. Creating and composing XML documents 221

Figure 8-22 Mappings table shows the mapping between the source and target lists

3. After you map the Customer target list, you can map the columns under it. To
map each of the target columns, select the corresponding row in the Source
column. From the suggestion list, select the appropriate item to map. The item
recommendations in the suggestion list are based on the name and data type
of the Target column.

After mapping all the columns, the Mappings table is complete, as shown in
Figure 8-23 on page 223.

222 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-23 The complete Mappings table

Automatically mapping the columns

One bank schema contains over a hundred columns. It is tedious to map each
column in the target to the corresponding column in the source. The Composer
step provides an easy mechanism to map multiple columns simultaneously by
using the Auto Map function. Follow these steps to use the Auto Map feature:
1. Map the source list DSLink2 to the target list Customer.
2. Select the row corresponding to the target list Customer.
3. Click Auto Map. All the columns under the list Customer are automatically
mapped.

The mapping is based on the similarity between the name and data type of the
source and target columns. The automatic mapping suggestion might not always
be correct. The Auto Map feature also cannot map columns of different data
types. Therefore, if the target column is an integer and all source columns are of

Chapter 8. Creating and composing XML documents 223

 type string, then no mapping occurs for the integer column by using the Auto Map
feature.

Selecting mapping candidates by using the More option

When the output of the Input step contains many columns, the suggestion list
cannot list all these columns. By default, the suggestion list shows only five
source items. If the required column is not available in the suggestion list, you
can use the More option. The More option shows the entire source tree structure
from which the source column to map can be selected. To use the More option,
follow these steps:
1. Select the row in the Source column that you need to map to the Target
column.
2. The More option is available near the end of the suggestion list, as shown in
Figure 8-24. Select More.

Figure 8-24 The suggestion list that shows the More option

224 InfoSphere DataStage for Enterprise XML Data Integration

3. The Target window opens, as shown in Figure 8-25. In this window, you can
see the entire schema tree.

Figure 8-25 The Target window

4. Select the item to map and click OK.

The mapping is complete.

Mapping constants to a target column

The bank needs to create an XML file. In this file, the Account_Type needs to
always be set to Savings, no matter what type of data comes from the database
columns. Therefore, a constant value needs to be assigned to the output column
Account_Type. Use the Composer Step Constant option. To map the target item
Account_Type to the constant value Savings, follow these steps:
1. Ensure that the parent list of the target item is mapped. Therefore, to assign a
constant value to the item Account_Type, Customer needs to be mapped.
Map DSLink2 to the list Customer.
2. Select the row in the source corresponding to the target item Account_Type.
The suggestion list is shown in Figure 8-26 on page 226.

Chapter 8. Creating and composing XML documents 225

Figure 8-26 The suggestion list with the Constant option

3. The last option in the suggestion list is Constant. Select the Constant option
to open the Constant window. In the Constant window shown in Figure 8-27,
you can enter the value to assign to the target item. To assign the value
Savings, enter Savings in the Constant window and click OK.

Figure 8-27 The Constant Window

226 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 8-28 shows a constant value assigned to the target column
Account_Type.

Figure 8-28 The Constant value Savings assigned to the target item Account_Type

8.1.5 The Header tab

The bank requires comments on the XML document. These comments are used
by the application that creates the customer reports. Use the Header tab in the
Composer Step to add comments to the XML document that is created by the
Composer Step. Select the Include Comments option to add comments. The
text box for Comments becomes active. You can enter your comment in the text
box, as shown in Figure 8-29 on page 228.

Chapter 8. Creating and composing XML documents 227

Figure 8-29 The Header tab in the Composer Step

You do not need to enter the comments in the format <!-- -->. The tags are
automatically added by the XML Composer step while you add comments.

You can also add Processing Instructions to the output XML file by choosing the
Include Processing Instructions option. The Processing Instruction must be
added between the tags <? ?>, as shown in Figure 8-30 on page 229.

228 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-30 The Header tab: Comments and processing instructions are added

The Include XML Declaration is selected by default in the Header tab. Use this
option to add the XML declaration to the output XML file:

<?xml version="1.0" encoding="UTF-8">

When Include XML Declaration is not selected, the output XML file does not
contain the XML declaration.

Use the Generate XML Fragment option when you do not want to include the
XML declaration, comments, and processing instructions in the output XML
document.

Chapter 8. Creating and composing XML documents 229

8.1.6 The Format tab
The bank requires a formatted output XML file. To add formatting, use the Format
tab. In the Format tab, select Format Style. The formatting options are enabled,
as shown in Figure 8-31.

Figure 8-31 The Format tab of the Composer Step

The Indentation Length helps specify the degree of indentation that is required.
The New Line Style helps specify the new line character to use within the file.
Formatting affects performance.

8.1.7 The output of the Composer step

To configure the Composer step, you need to configure the XML Target,
Document Root, Mappings, Header, and Format tabs. The output of the
Composer Step is seen in the Output tab. The output of the Composer Step is in

230 InfoSphere DataStage for Enterprise XML Data Integration

 the group XML_Composer:result. The output depends on the option that is
selected in the XML Target tab.

When the Write to File option is selected in the XML Target tab, the output XML
file is created by the Composer step. Therefore, a single item called
output-filename shows under the group XML_Composer:result, as shown in
Figure 8-32. This item contains the name of the output XML file in the Filename
Prefix text box in the XML Target tab.

Figure 8-32 Composer step output for the Write to File option

When the Pass as String or Pass as Large Object option is selected, the XML
Composer step does not write the file directly to the server. The XML file is
passed to a stage after XML Stage. Therefore, the item result-string is under the
group XML_Composer:result, as shown in Figure 8-33. This single item holds the
entire XML file.

Figure 8-33 Composer step output for the Pass as String or Pass as Large Object option

8.2 Creating the hierarchy

The XML file that we created by using the Composer step in the previous
sections contains a list called Customer. Therefore, you can directly map the
output of the Input step to the Composer step. But if the XML schema contains
multiple lists, the output of the Input step cannot directly be mapped to the
structure in the Composer step. Figure 8-34 on page 232 shows an XML schema
structure that contains multiple lists. To map the address and phoneNumber
elements, the parent list address_info needs to be mapped first. You cannot map
the input list DSLink2 to address_info, because it is already mapped to the list
Customer.

Chapter 8. Creating and composing XML documents 231

 Figure 8-34 XML schema

To create an XML file based on the structure shown in Figure 8-34, the same
hierarchy must be created so that the mapping is possible. The hierarchy can be
created by using the Regroup step or the HJoin step in the assembly. We explain
these steps in 8.2.1, “The Regroup step” on page 232, and 8.2.2, “The HJoin
step” on page 240.

8.2.1 The Regroup step

The Regroup step is required to create a hierarchical structure. It converts a
single list to two lists with a parent-child relationship. During this process, the
Regroup step also removes any redundancy in the incoming data. For example,
consider the input that comes from the database that contains the department
code and the employee information that are shown in Table 8-1. By using the
Regroup step a hierarchy can be created in which the department code is set as
the parent and the employee information is set as the child. The Regroup step
also removes the redundancy in the Department ID column by placing the
employees Jane Smith and John Doe1 under the same department.

Table 8-1 Removing redundant data elements

Department ID Employee name Employee ID

A100 Jane Smith M098

A100 John Doe1 M765

B200 John Doe2 U789

The bank has a schema that contains multiple lists, as shown in Figure 8-34. To
compose an XML file with this structure, use the Composer step. But, to be able

232 InfoSphere DataStage for Enterprise XML Data Integration

 to map the data from the Input step to the Composer step, you need to create a
hierarchy. You can use the Regroup step to create the hierarchy.

To configure the Regroup step, it needs to be added to the assembly. To add the
Regroup step, use the following steps:
1. Open Assembly Editor and click Palette in the Assembly Outline bar. You
can see the various steps in the Palette, as shown in Figure 8-35.

Figure 8-35 The Palette

Chapter 8. Creating and composing XML documents 233

 2. Drag the Regroup Step from the Palette onto the Assembly Outline bar after
the Input Step. The Regroup Step is added between the Input Step and the
Composer Step, as shown in Figure 8-36.

Figure 8-36 The Assembly Editor

To configure the Regroup step, follow these steps:

1. Set List to Regroup to the input list DSLink2. The list DSLink2 is selected,
because it contains all the parent and child items under it.
2. Set the Scope to top. The scope defines the location of the output of the
Regroup step. When top is defined as the scope, the output of the Regroup
step is under the list top. The list that is selected as the scope must always be
the parent of the list that is selected as the List to Regroup.
3. After you perform steps 1 and 2, all the items under the list DSLink2 can be
seen in the Child Items column.
4. Because the items Account_Number, Account_Holder_Name,
Account_Holder_DOB, and Account_Type are under the parent list Customer,
these items are dragged into the Parent Items column.

234 InfoSphere DataStage for Enterprise XML Data Integration

 5. The parent item Account_Number determines which child item is under it.
Therefore, Account_Number must be set as the key in the Keys tab of the
Regroup Step. Only the parent items can be selected as key values.

The Regroup Step is configured, as shown in Figure 8-37 and Figure 8-38 on
page 236.

Figure 8-37 The Regroup Step

Chapter 8. Creating and composing XML documents 235

Figure 8-38 Keys tab of the Regroup Step

The input and output of the Regroup step is shown in Figure 8-39 on page 237.
The output of the Regroup step is under the list Regroup:result. Directly under
this list are the parent items. A child list, DSLink2, is also under Regroup:result.
The name of the child list is same name as the List to Regroup, which we
configured in the Regroup step. As shown in Figure 8-39 on page 237, the child
list DSLink2 contains all the child items.

236 InfoSphere DataStage for Enterprise XML Data Integration

 Inp ut to Regroup Step

O utput of Regro up Step

Figure 8-39 Input and output of the Regroup step

The Output of the Regroup step can now be mapped to the XML schema in the
Mappings tab of the Composer step, as shown in Figure 8-40.

Figure 8-40 Mapping between Regroup output and Composer step

Regroup lists are clustered by key-optimize execution

When the input data is stored in a database, most often, the data is sorted. When
this data is fed to the Regroup step, the Regroup step searches for all the unique
key values and rearranges the data into a hierarchy. If the input data is already
presorted based on the key value defined in the Regroup step, it is easy for the
Regroup step to create the parent-child hierarchy. With the parent-child hierarchy,
the step does not need to search through the entire input data for the unique key
values.

The “Input records of regroup list are clustered by key - optimize execution”
option in the Keys tab instructs the Regroup Step to regroup in a streaming

Chapter 8. Creating and composing XML documents 237

 fashion without storing the entire list in memory. This option also enhances the
performance of the step and the overall performance of the XML Stage. This
option is depicted in Figure 8-41.

Figure 8-41 The Input records of regroup lists are clustered by key - optimize execution option

In case the data is not sorted in the database, the data can be sorted before
feeding it to the XML Stage. We advise that you sort in DataStage rather than in
XML Stage, because the DataStage Sort operates in parallel and can easily sort
the relational data that comes from the database. The Sort step within the XML
Stage is appropriate for sorting hierarchical data. Figure 8-42 shows the overall
DataStage job design. The option “Input records of regroup lists are clustered by
key - optimize execution” is used within the Regroup step.

Figure 8-42 The DataStage job design

238 InfoSphere DataStage for Enterprise XML Data Integration

Creating multiple levels of hierarchy
The bank also uses complicated schemas that involve multiple levels of
hierarchy. Figure 8-43 shows a complex schema; the parent list Customer
contains the list address_info, and address_info contains the list phoneNumber.

Figure 8-43 An XML schema with multiple, hierarchical levels

To create this hierarchy, two Regroup steps can be used. In the first Regroup
step, create the first level of hierarchy, which is between Customer and
address_info. The address_info link also contains the phoneNumber columns, as
shown in Figure 8-44.

Figure 8-44 The hierarchy created by the first Regroup step

The second Regroup step creates the second hierarchy level, which is between
address_info and phoneNumber. Remember to set the scope in the second

Chapter 8. Creating and composing XML documents 239

 Regroup step to Customer so that the output of the second Regroup step is
under the list Customer. Figure 8-45 shows the final hierarchy that is created.

Figure 8-45 Final hierarchy created by using two Regroup steps

The address_info list can be ignored, because the actual list is

address_info_final.

8.2.2 The HJoin step

The bank must create an XML file from source data in two DB2 tables. The data
in the databases must be combined to create the XML file. Therefore, two input
links are necessary to the XML Stage. Figure 8-46 on page 241 shows the
DataStage job design.

240 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-46 The HJoin DataStage job design

The XML Stage needs to combine the data from the two input links into a single
XML structure, as shown in Figure 8-47.

Figure 8-47 Combining two input files to create a single XML file

The HJoin step is used to join data that comes from two input links. This join is
called a Hierarchical Join, because the step creates a hierarchical structure.
Therefore, the HJoin step joins the data and creates the hierarchy. To be able to
map the data from the input database to the Composer step, the HJoin step must
be used.

The HJoin step must be added to the assembly to configure the HJoin step.
Follow these steps to add the HJoin step to the assembly:
1. Open Assembly Editor and click Palette in the Assembly Outline bar. The
steps show in the Palette, as shown in Figure 8-48 on page 242.

Chapter 8. Creating and composing XML documents 241

Figure 8-48 The Palette in the Assembly Editor

2. Drag the HJoin step from the Palette onto the Assembly Outline bar after the
Input Step. The HJoin step is added between the Input Step and the
Composer Step, as shown in Figure 8-49 on page 243.

242 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-49 The HJoin Step added to the Assembly Outline

To configure the HJoin step, follow these steps:

1. For the Parent List and Child List options, specify the input lists, DSLink2 and
DSLink4, to the HJoin step. Because the items under list DSLink2 are the
parent items in the XML schema structure defined in the Composer step,
specify DSLink2 as the Parent List. Specify DSLink4 as the Child List.
2. Both the database tables have the common column, Account_Type, which
helps to create a relationship between the two tables. This column is used to
join the data that comes into the HJoin step. Therefore, select Account_Type
for Parent Keys and Child Keys.
3. The Optimization Type determines where the join occurs. For large input data,
use Disk-based, which is the default, for the Optimization Type.

The HJoin step is configured, as shown in Figure 8-50 on page 244.

Chapter 8. Creating and composing XML documents 243

Figure 8-50 The complete HJoin Step

Figure 8-51 on page 245 shows the output of the HJoin step. The output contains
a nested structure. Under the input parent list DSLink2, the child list is added
along with the prefix HJoin, for example, HJoin:DSLink4.

244 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-51 The output of the HJoin step

In several rows of the child list, the value of the column Account_Type might not
match a value in the parent list. All of these rows are placed under a single group
called HJoin:orphans, as shown in Figure 8-52.

Figure 8-52 The orphan child records

The output of the HJoin step can now be mapped to the XML schema in the
Mappings tab of the Composer step, as shown in Figure 8-53 on page 246.

Chapter 8. Creating and composing XML documents 245

 Figure 8-53 The mapping

8.2.3 Regroup step versus HJoin step

You can use both the Regroup step and the HJoin step to create a hierarchical
structure.

Use the Regroup step when a single input file is provided to the XML Stage, and
the input relational data has duplicate rows. For example, consider the data that
is shown in Table 8-2.

Table 8-2 Single input file with duplicate rows

Department Employee Employee Date of birth Employee ID
ID first name last name

A100 Jane Smith 1986-08-06 Z1098

B200 John Doe1 1977-10-30 Y9876

A100 John Doe2 1969-01-01 X7656

C300 Jane Smith1 1988-01-09 W2345

C300 Jane Smith2 1972-05-26 T7400

To create a hierarchical structure of each department and its employees, you can
use the Regroup step. Several of the Department IDs are repeated. By using the
Regroup step, the redundancy in the Department IDs can be removed.

246 InfoSphere DataStage for Enterprise XML Data Integration

Use the HJoin step when multiple input sources are required. This step joins the
data in the input source to create a hierarchical structure. Consider the two input
sources that are shown in Table 8-3 and Table 8-4.

Table 8-3 Multiple sources: Input 1

Department ID Department name

A100 Human resources

B200 Product sales

C300 Software development

Table 8-4 Multiple sources: Input2

Department Employee Employee Date of birth Employee ID
ID first name last name

A100 Jane Smith 1986-08-06 Z1098

B200 John Doe1 1977-10-30 Y9876

A100 John Doe2 1969-01-01 X7656

C300 Jane Smith1 1988-01-09 W2345

C300 Jane Smith2 1972-05-26 T7400

When an HJoin is performed between these tables, the Department IDs A100
and C300 repeat, because the HJoin step does not remove any redundancy
while performing a join. Therefore, the Regroup step must be used. Because the
Regroup step requires a single input list, both tables can be combined by using
the Join_Stage that is available in DataStage. Figure 8-54 shows the DataStage
job design.

Figure 8-54 The Join DataStage Job design

Chapter 8. Creating and composing XML documents 247

 To create a hierarchy with multiple levels, you can use multiple Regroup steps or
HJoin steps in the assembly.

8.2.4 H-Pivot step

The H-Pivot step is used to transform multiple columns into a single hierarchical
column under a list. For example, consider the input from the database, which is
information about a departmental store. In the input, two columns, BeautyPr and
HomePr, contain information about the products in the store. Both the columns,
BeautyPr and HomePr, need to be mapped to a single column that is under a list,
as shown in Figure 8-55. Use the H-Pivot step to map multiple columns in the
input to a single column under a list in the output.

BeautyPr
HomePr

Figure 8-55 Mapping input columns to the schema structure

The bank uses a schema that contains the customer account information, as
shown in Figure 8-56.

Figure 8-56 The XML schema structure

The Composer step is used to compose XML files. To map the data from the
Input step to the Composer step, the hierarchy needs to be created. The input
columns are shown in Figure 8-57 on page 249. Therefore, two input columns,

248 InfoSphere DataStage for Enterprise XML Data Integration

phoneNumber1 and phoneNumber2, need to be mapped to a single output
column, text(), under the list phoneNumbers.

Figure 8-57 The input columns

You need to add the H-Pivot step to the assembly before you configure it. To add
the H-Pivot step, use the following steps:
1. Open Assembly Editor and click Palette in the Assembly Outline bar, which
shows the various steps in the Palette, as shown in Figure 8-58 on page 250.

Chapter 8. Creating and composing XML documents 249

Figure 8-58 The Palette in the Assembly Editor

2. Drag the H-Pivot step from the Palette onto the Assembly Outline bar after
the Input Step. The H-Pivot Step is added between the Input Step and the
Composer Step, as shown in Figure 8-59 on page 251.

250 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-59 The H-Pivot Step in the Assembly Editor

To configure the H-Pivot step, use the following steps:

1. Set the Scope as the list that contains the phoneNumber1 and
phoneNumber2 fields. Because these fields are under the input list DSLink2,
select DSLink2 for Scope.
2. In the Columns drop-down list, select the input columns. Select
phoneNumber1 in the first drop-down list and phoneNumber2 in the second
drop-down list. Click OK.

The H-Pivot step is configured, as shown in Figure 8-60 on page 252.

Chapter 8. Creating and composing XML documents 251

Figure 8-60 Configured H-Pivot Step

The output of the H-Pivot step is contained under the scope DSLink2. The output
of the H-Pivot step is under the group H-Pivot:result, as shown in Figure 8-61 on
page 253. It contains the rows list, which contains the child items: name and
value. Name contains the name of the fields selected as columns (in this
scenario, Name stores phoneNumber1 and phoneNumber2). Value contains the
data that is stored in the phoneNumber1 and phoneNumber2 columns.

252 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 8-61 The output of the H-Pivot step

The output of the H-Pivot step can now be mapped to the phoneNumbers list, as
shown in Figure 8-62.

Figure 8-62 Mapping the H-Pivot step output and Composer step schema structure

8.3 Multiple XML documents: Single Composer step

Use the Mappings tab of the Composer step to map the incoming data to the
XML schema structure to create the XML document. The document_collection is
the first target list in the Mappings table. Based on the mapping to this list, one or
multiple XML documents are created. By default, the top list is mapped to the
document_collection list so that the XML Composer step creates only one XML
file.

Chapter 8. Creating and composing XML documents 253

 The bank must create XML files by using the data in the database. The bank
wants to create one XML file for every input row that comes from the DB2
database table. The input structure must be mapped to the schema structure in
the Mappings tab of the Composer step. To create one XML file for every
incoming input row, map the input list DSLink2 to the document_collection list.
Now, you can map the rest of the target items, as shown in Figure 8-63.

Figure 8-63 Mapping the complete document_collection list

The XML Composer result is contained in the top list when the top list is mapped
to the document_collection list. When the input list DSLink2 is mapped to the
document_collection list, the XML Composer result is contained in the DSLink2
list, as shown in Figure 8-64.

Figure 8-64 The XML Composer result that is contained in the DSLink2 list

254 InfoSphere DataStage for Enterprise XML Data Integration

8.3.1 Target tab configuration for creating multiple XML documents
The XML Target tab describes the method of writing the XML document to the
target. It provides three options for writing the XML file: Write to File, Pass as
String, and Pass as Large Object. We described these options in detail in 8.1.1,
“Writing XML files” on page 202. The configuration of the XML Target tab also
determines how to write the multiple XML files, which we created, to the target. In
the Mappings tab, the input list DSLink2 is mapped to the document_collection
list. Therefore, for every input row under the list DSLink2, a new XML file is
created by the Composer step.

When the Write to File option is selected in the XML Target tab of the Composer
step, the Composer step directly writes the XML file onto the server. It creates
one XML document for every input row. For example, if three input rows come
into the XML Stage and the Filename Prefix value is file, the Composer step
creates three files with the names file.xml, file_1.xml, and file_2.xml.

When the Pass as String or Pass as Large Object option is selected in the XML
Target tab of the Composer step, the Composer step passes the created XML file
to the stage after XML Stage. The Composer step does not directly write the file
onto the server. By choosing this option, the Composer step creates a single item
that contains both files. For example, if there are two input rows, the result of the
Composer step is a single item, result_string, which contains the data that is
shown in Example 8-3.

Example 8-3 Single item for multiple rows

<?xml version="1.0" encoding="UTF-8"?>
<Account>
<Customer>
<Account_Number>C198765487</Account_Number>
<Account_Holder_Name>John Doe1</Account_Holder_Name>
<Account_Holder_DOB>02-07-1980</Account_Holder_DOB>
<Account_Holder_address>San Jose,CA</Account_Holder_address>
<Account_Holder_phoneNumber>123-456-654-4567</Account_Holder_phon
eNumber>
<Account_Type>Private</Account_Type>
</Customer>
</Account>
<?xml version="1.0" encoding="UTF-8"?>
<Account>
<Customer>
<Account_Number>C198765444</Account_Number>
<Account_Holder_Name>Jane Smith</Account_Holder_Name>
<Account_Holder_DOB>1975-01-01</Account_Holder_DOB>
<Account_Holder_address>San Jose,CA</Account_Holder_address>

Chapter 8. Creating and composing XML documents 255

 <Account_Holder_phoneNumber>123-789-987-2457</Account_Holder_phon
eNumber>
<Account_Type>non-savings</Account_Type>
</Customer>
</Account>

Both the XML documents are embedded in a single file.

8.4 Configuring the Output step

When the Pass as String or Pass as Large Object option is selected in the XML
Target tab of the Composer step, the Composer step passes the created XML file
to the stage after the XML Stage. The Output step must be used to pass the
created XML file to the downstream stage. For example, if the XML file must be
written into a pureXML column of a database, the Output step is required.

When an output link is provided from the XML Stage, the link is a list within the
assembly. The output columns are items in the assembly. The columns for each
of the output links are visible in the Output tab of the Output Step. The Mappings
tab is used to map the output of the Composer Step to the columns of the
database. As shown in Figure 8-65 on page 257, the output link DSLink3 has a
single column result to which the Composer Result must be mapped.

256 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-65 The Mappings tab for mapping the output of the Composer Step

When the top list is mapped to the document_collection list, the output of the
Composer step is under the list top. The Composer result is contained in a single
item called result_string. To map result_string to the output column, you must first
map the parent list DSLink3. Because result_string is under the list top, top is
mapped to the target list DSLink3. Then, the source item result_string is mapped
to the target item result. Figure 8-66 on page 258 shows the complete mapping.

Chapter 8. Creating and composing XML documents 257

Figure 8-66 Mapping the parent list to result_string

If the input link DSLink2 is mapped to the document_collection item in the

Mappings tab of the Composer step, the output of the Composer step is under
the input list DSLink2. To map result_string to the output column, you first must
map the parent list DSLink3. Because the item result_string is under the input list
DSLink2, DSLink2 is mapped to the target list DSLink3. Then, the source item
result_string is mapped to the target item result. Figure 8-67 on page 259 shows
the complete mapping.

258 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-67 The complete mapping for result_string

8.5 Creating XML files for various kinds of schemas

When the XML schema is imported into the Schema Library Manager, it is
converted into a simplified model. This simplified model is used by the Composer
step to create the XML files. We describe the representation of the schema in the
assembly in detail in Chapter 6, “Introduction to XML Stage” on page 131. For
certain schema definitions, the XML Stage adds its own items to the schema
representation. You must pay special attention to composing the XML files for
these schemas.

8.5.1 Creating an XML file for a schema with a choice definition

The bank uses a schema in which one of the complex elements is defined with
xs:choice. The XML schema is shown in Figure 8-68 on page 260.

Chapter 8. Creating and composing XML documents 259

 The item @@choiceDiscriminator is added by the XML Stage before the
self_employed and hired elements to signify the xs:choice in the schema
definition. The @@choiceDiscriminator holds the name of the choice branch
selected at run time. For the XML schema that is shown in Figure 8-68, the
output XML file can contain either the element self_employed or the element
hired, but not both elements. While composing the XML file for this schema, the
correct value needs to be mapped to the @@choiceDiscriminator.

Figure 8-68 The item @@choiceDiscriminator mapped

These mapping rules apply when you map the @@choiceDiscriminator item:
򐂰 If the input data is present only for the element self_employed, the constant
value self_employed must be mapped to the target item
@@choiceDiscriminator, as shown in Figure 8-69 on page 261. The constant
value must be the same as the name of the choice item.

260 InfoSphere DataStage for Enterprise XML Data Integration

Figure 8-69 The mapping rules while mapping @@choiceDiscriminator item

򐂰 If the input data is present only for the element hired, the constant value hired
must be mapped to the target item @@choiceDiscriminator, as shown in
Figure 8-70.

Figure 8-70 A constant value mapped

򐂰 Sometimes the incoming data can have both elements, that is, one row is
self_employed and the other row is hired. The constant value cannot be

Chapter 8. Creating and composing XML documents 261

 mapped to the item @@choiceDiscriminator, because, during design time,
you cannot verify which element is present in the input. In this case, a
Transformer stage is used before the XML Stage. In the Transformer stage, a
new column is added called choice. The value of this column is determined by
using “If...else” conditioning. The following example shows the condition in the
Transformer stage:
If IsNotNull(DSLink2.self_employed) Then "self_employed" Else
"hired"
The column choice is then mapped to the @@choiceDiscriminator item in the
Mappings tab of the Composer step, as shown in Figure 8-71.

Figure 8-71 The column choice mapped to the @@choiceDiscriminator item

At run time, if the input row contains data for the self_employed element,
choice contains the value self_employed. The Composer step knows which
element must be present in the output XML file.
򐂰 XML schemas might use choice elements with the tns extension. This
extension means that the self_employed element is present in the
namespace that is represented by tns. When mapping the value to the
@@choiceDiscriminator, you need to take care of the namespace, too.
Therefore, if the input has the data for the self_employed element and the
namespace is http://check123, you map the constant value
“{http://check123}self_employed” to the @@choiceDiscriminator item.

262 InfoSphere DataStage for Enterprise XML Data Integration

8.5.2 Creating the XML file for a schema with type derivatives
The bank uses a schema in which an element is derived from another element.
Figure 8-72 shows the XML schema and its representation in the assembly. The
element office_address in the schema is derived from the complex type
home_address. Because office_address is derived from home_address, the
element address can also be the office_address type. Therefore, the item
@@type is added under the address group, so that at run time, the @@type
item can determine which structure to use.

Figure 8-72 The item @@type for handling type derivatives

The following mapping rules apply when you create an XML file for a schema
with @@type:
򐂰 If nothing is mapped to the @@type item, the output XML file contains only
the elements of home_address. As you can see in Figure 8-73 on page 264,
the mapping occurred for home and office address fields, but because
nothing is mapped to @@type, the output contains only home_address.

Chapter 8. Creating and composing XML documents 263

 Figure 8-73 Nothing is mapped to @@type item

򐂰 If the constant value home_address is mapped to the @@type,

home_address is in the output XML file.
򐂰 If the constant value office_address is mapped to the @@type,
office_address is in the output XML file.
򐂰 In case the input data can have either office or home address, a Transformer
stage can be used before the XML Stage. In the Transformer stage, the value
for a new column type can be determined by using “If...else” conditioning. The
column type can then be mapped to the @@type item in the Composer step.

264 InfoSphere DataStage for Enterprise XML Data Integration

 9

Chapter 9. Transforming XML

documents
The XML Stage can read, write, and transform XML files. The Parser step is used
for reading XML files. The Composer step is used for writing or creating XML
files. Sometimes, you are required to perform certain operations on the XML
data, for example, sorting the XML file or filtering the XML content to remove
unwanted data. The XML Stage provides steps to transform XML files. We
describe these steps in detail in this chapter.

We use the same XML structure to explain each step. The XML file contains the
details of a bank account. Figure 9-1 on page 266 shows the XML structure in
the assembly.

© Copyright IBM Corp. 2012. All rights reserved. 265

 Figure 9-1 The XML schema structure in the Assembly

266 InfoSphere DataStage for Enterprise XML Data Integration

9.1 The Sort step
The Sort step is used to sort an item under a list in ascending or descending
order. The Sort step works in a similar manner to the Sort stage in DataStage.
The Sort stage sorts the incoming data in a link based on a key. Similarly, the
Sort step sorts a list in an XML file based on an item selected as the key.

For the XML schema structure that is shown in Figure 9-1 on page 266, the
Account information needs to be sorted in ascending order based on the
FirstName field. To sort the XML file, first the XML file needs to be read by using
the Parser step. The Sort step must be added after the Parser step to sort the
content of the XML file. Follow these steps to configure the Sort step that is
depicted in Figure 9-2 on page 268:
1. The List to Sort specifies the list that needs to be sorted in the XML file. In the
List to Sort drop-down list, select the list Account.
2. The Scope in the Sort step defines the scope of the sort function. The
elements are sorted within the list that is selected as the Scope. The Scope
also determines the location of the output of the Sort step. The Scope must
always be the parent node of the list that is selected in the List to Sort field. In
this scenario, because the Account is selected as the List to Sort, only top
can be selected as the Scope. Select top. The output of the Sort step is now
under the list top. All the FirstName items that are under the top scope are
sorted.
3. The XML file must be sorted on the FirstName item. Therefore, for Keys,
select FirstName from the drop-down list.
4. You can select the sort order in the drop-down list under Order. The sorting
options are to sort in Ascending or Descending order. Select Ascending.

The configuration of the Sort Step is shown in Figure 9-2.

Chapter 9. Transforming XML documents 267

 Figure 9-2 The configuration of the Sort Step

Because we selected top as the Scope, the output of the Sort step is under top,
as shown in Figure 9-3 on page 269. The output of the Sort step is in a group
called Sort:result. The output shows the list Account sorted in ascending order
based on the FirstName field. This output can then be mapped to the Composer
step to create the sorted XML file.

268 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 9-3 The output of the Sort step

Sorting has special considerations. See 10.5.1, “Sort step” on page 318 for the
details.

9.2 The Aggregate step

The Aggregate step performs hierarchical aggregation on an item in a list. It is
similar to the Aggregate stage in the DataStage Designer Canvas. The
Aggregate stage groups data for an input link and performs count and sum
functions. Similarly, the Aggregate step in XML Stage groups data within a list
and performs functions, such as Count and Concat.

A customer at our scenario bank has either a loan account or a deposit account.
You can use the Aggregate step to count the number of customers based on the
type of account. To count the number of accounts of a particular account type in
the XML file, first the XML file needs to be read by using the Parser step. The
Aggregate step must to be added after the Parser step to aggregate the
hierarchical data.

Chapter 9. Transforming XML documents 269

 Follow these steps to configure the Aggregate step, as depicted in Figure 9-6 on
page 272:
1. The List to Aggregate field specifies the list whose items need to be
aggregated in the XML file. In the current scenario, the aggregation can be
performed by using any of the items. We use the item called FirstName. In the
List to Aggregate drop-down list, select the list that contains the FirstName
field. Because FirstName is under the list Account, select Account in the List
to Aggregate drop-down list.
2. The Scope field in the Aggregate step defines the scope of the Aggregate
function. The elements are aggregated within the list that is selected as the
scope. The scope also determines the location of the output of the Aggregate
step. The scope must always be the parent node of the list that is selected in
the List to Aggregate field. In this scenario, because Account is selected for
the List to Aggregate, only top can be selected as the scope. Select top.
Because the Scope is set as top and the item selected for aggregation is
FirstName, the Aggregate step checks for all the FirstName fields within the
top list and then aggregates them. Therefore, this step looks for FirstName in
the input file and aggregates it.
3. In the Items columns, you select the item that needs to be grouped. From the
drop-down list, select the FirstName field.
4. In the Functions column, select the function that you want to perform on the
item FirstName. Multiple functions are available for each of the data types. In
the current scenario, because we need to identify the number of accounts, we
select the Count function from the drop-down list.
5. The Aggregate step is configured. The output of the Aggregate step is directly
under top, which we selected as the scope. The output is in a group called
Aggregate:result, as shown in Figure 9-4. All the FirstName fields in the XML
file are counted without considering the account types.

Figure 9-4 The output of the Aggregate step

6. Next, because the aggregation is based on the account_type field, we select

account_type from in the Aggregation Keys drop-down list. The output of the
Aggregate step is a list called Aggregate:result, which is under top, as shown
in Figure 9-5 on page 271. The Aggregate:result is a list, because one output
exists for each account_type value. The output also contains the

270 InfoSphere DataStage for Enterprise XML Data Integration

 account_type field. For example, if an XML file includes 10 accounts, and four
accounts are loan account_type accounts and the rest of the accounts are
deposit account_type accounts, the output of the Aggregate step looks like
Example 9-1.

Example 9-1 The output of the Aggregate step

4,loan
6,deposit

Figure 9-5 The output of the Aggregate step when a key value is specified

The configuration of the Aggregate Step is shown in Figure 9-6 on page 272.

Chapter 9. Transforming XML documents 271

 Figure 9-6 The configuration of the Aggregate Step

You can map the output of the Aggregate step to various columns of a Sequential
file by using the Output step.

Appendix B, “Transformer functions” on page 369 lists the available functions in

the Aggregate step.

9.3 The Switch step

The Switch step is used to filter the XML file and classify the items into one or
more targets. The filtering is based on constraints that you specify. Each
constraint is associated with an output list. A default list contains the items that
do not satisfy any of the constraints.

A customer at our scenario bank has either a loan account or a deposit account.
You can use the Switch step to write the account information for each type into a
separate output file. The XML file must be read by using the Parser step. You
must add the Switch step after the Parser step to filter the items in the XML file.
Follow these steps to configure the Switch step, as depicted in Figure 9-9 on
page 275:

272 InfoSphere DataStage for Enterprise XML Data Integration

1. The List to Categorize field is the list that contains the item that is used to
specify the constraint. The item account_type is under the list Account.
Therefore, select Account from the drop-down list for the List to Categorize
field.
2. The Scope defines the scope of the Switch step. The elements are filtered
within the list that is selected as the scope. The scope also determines the
location of the output of the Switch step. The scope must always be the
parent node of the list that is selected in the List to Categorize field. In this
scenario, because we selected Account for the List to Categorize field, only
top can be selected as the scope. Select top. Because the scope is set to top
and the item for filtering is account_type, the Switch step checks for all the
account types within the top list and then filters the list. Therefore, the entire
input file is filtered.
3. To specify the constraint for filtering, add a target in the Switch step. To add a
target, follow these steps:
a. Click Add Target. The New Target Constraint window opens, as shown in
Figure 9-7.

Figure 9-7 The New Target Constraint window

b. Type loan for the Target Name. The Target Name is the name of the
constraint, which also is the name of the output list.
c. The Filter Field specifies the item on which the constraint is specified.
Select account_type from the drop-down list.
d. The Switch step provides multiple functions based on the type of filtering
that you want to perform. Select Compare in the drop-down list for
Function.
e. The account_type can be either loan or deposit. Therefore, either one of
these constant values can be specified as the constraint. For Parameter 1,

Chapter 9. Transforming XML documents 273

 select Constant and type loan to specify the string in the text box, as
shown in Figure 9-8.

Figure 9-8 The configuration of the New Target Constraint window

Figure 9-9 on page 275 shows the configuration of the Switch Step.

274 InfoSphere DataStage for Enterprise XML Data Integration

Figure 9-9 The configuration of the Switch Step

The output of the Switch step is in a group called Switch:filtered, as shown in

Figure 9-10 on page 276. Two lists are under this group: loan and default. The
loan list contains all the items that satisfy the constraint for the target loan. The
default list contains all the items that do not satisfy the loan constraint.

Chapter 9. Transforming XML documents 275

 Figure 9-10 The output of the Switch step

Each of the lists under Switch:filtered can be mapped to a separate output link by
using the Output step. Therefore, all the account information for the loan account
type is in one output file. All the account information for the deposit account type
is in a separate output file (the output file to which the default list is mapped).

See Appendix B, “Transformer functions” on page 369 for a list of functions that
are available in the Switch step.

9.4 The Union step

The Union step is used to combine two lists into a single output list. Two XML
files can be combined into a single file by using the Union step.

The XML files can be read by using two Parser steps. The Union step must be
used after the Parser steps to combine the two files.

276 InfoSphere DataStage for Enterprise XML Data Integration

Follow these steps to configure the Union step:
1. Specify the target schema in the Union Type tab. The target schema defines
the structure of the output of the Union step. The Union Type tab is similar to
the Document Root tab of the Parser and Composer steps. To specify the
Union Type, click Browse to select the schema element from the Schema
Library Manager. When you select the schema, it is defined in the Union Type
tab, as shown in Figure 9-11.

Figure 9-11 The Union Type tab of the Union Step

2. The Mappings tab is similar to the Mappings tab in the Composer step. You
can use the Mappings table to map the output of the Parser step to the target
structure in the Union step.
This Mappings table, as shown in Figure 9-12 on page 278, contains three
columns: Source, Result, and Target. The Source column shows the columns
that need to be mapped to the target items. The Result column shows the
status of the current mapping. The Target column contains two lists: the left

Chapter 9. Transforming XML documents 277

 list and the right list. Each list shows the items of the target structure. The
Target structure is the XML schema that is defined in the Union Type tab of
the Union Step.

Figure 9-12 The Mappings table in the Union Step

3. The left list is mapped to the output of the first Parser step, and the right list is
mapped to the output of the second Parser step, as shown in Figure 9-13 on
page 279.

278 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 9-13 The mapping between the Parser steps and the Union step

As shown in Figure 9-13, a source list needs to be mapped to the left and right
lists in the target. The Account lists under the first and second Parser steps must
be mapped to the Account fields under the left and right lists. To create lists on
the source side, you can use the Group to List projection. Follow these steps to
create the lists on the source side:
1. Go to the Output tab of the First Parser step. Right-click the
XML_Parser:result group, as shown in Figure 9-14 on page 280.

Chapter 9. Transforming XML documents 279

 Figure 9-14 The Group To List projection option

2. Select the Group To List projection. The group is converted to a list, as

shown in Figure 9-15 on page 281.

280 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 9-15 XML_Parser:result as a list

3. Go to the Output tab of the Second Parser step. Right-click the group
XML_Parser_1:result and select Group To List. The group is converted to a
list, as shown in Figure 9-16 on page 282.

Chapter 9. Transforming XML documents 281

 Figure 9-16 XML_Parser_1:result as a list

Follow these steps to map the data from the Parser steps to the Union step:
1. Double-click the Source column for the target list left. You see a drop-down
list, which is the suggestion list. The suggestion list shows all available source
lists that can be mapped to the target list. The absolute path,
top/XML_Parser:result, is shown in the suggestion list.
2. Select the source list top/XML_Parser:result from the suggestion list. When
you select top/XML_Parser:result, the mapping occurs for this row. A green
circle with a white check mark shows in the Result column, as seen in
Figure 9-17 on page 283.

282 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 9-17 Mapping the Source to the Target

3. Then, map top/XM_Parser:result /Bank/Account to the list Account, which

is under the left list.
4. After mapping the target list Account, you can map the columns under it. To
map each of the target columns, select the corresponding row in the Source
column. From the suggestion list, select the appropriate item to map. The item
recommendations in the suggestion list are based on the name and data type
of the Target column.
5. Follow steps 1 - 4 to map XML_Parser_1:result and its child items to the
right list.

Figure 9-18 on page 284 shows the Mappings table after mapping all the
columns.

Chapter 9. Transforming XML documents 283

 Figure 9-18 The Mappings table

The Union step is configured. The output of the Union step is in a list called
Union_result, as shown in Figure 9-19. The structure of the Union:result list is the
same structure that is defined in the Union Type. The output of the Union step
can be mapped to the Composer step to recreate the XML file. Therefore, two
XML files are combined into a single file.

Figure 9-19 The output of the Union step

284 InfoSphere DataStage for Enterprise XML Data Integration

9.5 The V-Pivot step
The V-Pivot step is used to transform the values in a column into a single row.

To demonstrate the use of the V-Pivot step, the Account schema is slightly
modified. The new schema representation is shown in Figure 9-20. The address
is a list that can contain the account holder home address and the account holder
office address. The address_type field is used to distinguish the addresses. It
contains office for office address and home for home address. When this
structure is parsed, it creates two output rows: one row for home address and
one row for office address. To map all the account holder information and the
address information in a single output row, you must use the V-Pivot step after
the Parser step.

Figure 9-20 XML schema for V-Pivot example

Follow these steps to configure the V-Pivot step that is depicted in Figure 9-21 on
page 287:
1. The Source of Rows field specifies the list that contains the field on which the
pivot is performed. In this scenario, the address_type field is unique for each
address list. Therefore, the pivot is performed on the address_type field.
Because the address_type field is under the list address, select address in
the drop-down list for the Source of Rows field.

Chapter 9. Transforming XML documents 285

 2. The Scope in the V-Pivot step defines the scope of the V-Pivot function. The
scope also determines the location of the output of the V-Pivot step. The
scope must always be the parent node of the list that is selected in the Source
of Rows field. For the V-Pivot step to run successfully, the pivoting field must
contain unique values. The address_type field contains two values: office and
home. These values are repeated for every Account. But, these values are
unique for a single account. Therefore, select Account for the Scope to
ensure that the pivoting is performed for one account at a time rather than for
all the accounts. The Scope in the V-Pivot step must be selected to ensure
that the values of the pivoting field are unique.
3. The Source of Column Names field specifies the pivoting field. Select
address_type from the drop-down list.
4. The Column Name field specifies the unique values that can be in the pivoting
column, for example, address_type. Because the address_type can be
office and home, enter these values in individual rows. The Column Name
field can accept alphabetical characters, numerals, and a few special
characters, such as the period (.), underscore (_), and hyphen (-). The column
names must start with either an alphabetical character or underscore. The
column names cannot contain special characters, such as the colon (:) and
the At sign (@).

The V-Pivot Step is now configured, as shown in Figure 9-21 on page 287.

286 InfoSphere DataStage for Enterprise XML Data Integration

Figure 9-21 The configuration of the V-Pivot Step

The output of the V-Pivot step is in a group called V-Pivot:result, which contains
two nodes: office and home. These names are based on the values that are
specified for the Column Names. The home node holds the house address, and
the office node holds the office address. The output of the V-Pivot step is under
the Account list, because the Scope is set to Account. See Figure 9-22 on
page 288.

Chapter 9. Transforming XML documents 287

 Figure 9-22 The output of the V-Pivot step

You can map the output of the V-Pivot step to a sequential file by using the
Output step.

9.6 The OrderJoin step

The OrderJoin step is used to join two items in a list based on their positions.

A hospital stores its patient chart information in two XML files. One XML file
stores information about the doctor who treated the patients. The other XML file
stores the medications that are recommended by the doctor. We must merge
both sets of details into one XML file, as shown in Figure 9-23 on page 289. We
are not required to combine the exact medication that is prescribed with the
doctor who prescribed it. The information merely needs to be merged so that the
doctor information is followed by all the prescribed medicines. The OrderJoin
step can be used to create this XML file.

288 InfoSphere DataStage for Enterprise XML Data Integration

 <?xml version="1.0" encoding="UTF-8"?>
<Hospital>
<Provider>
<doctorID>Z1234</doctorID>
</Provider>
</Hospital> <?xml version="1.0" encoding="UTF-8"?>
<Hospital>
<Provider>
<doctorID>Z1234</doctorID>
</Provider>
<Medication>
<drugName>Aspirin</drugName>
</Medication>
<?xml version="1.0" encoding="UTF-8"?> <Medication>
<Hospital> <drugName>Dcold</drugName>
<Medication> </Medication>
<dr ugName>Aspirin</drugName> </Hospital>
</Medication>
<Medication>
<dr ugName>Dcold</drugName>
</Medication>
</Hospital>

Figure 9-23 The output of the OrderJoin step

In this scenario, two Parser steps are required to read each of the XML files. The
XML schema structure of each of the XML files is shown in Figure 9-24. The
OrderJoin step is added after the Parser steps.

Figure 9-24 The schema structures of the input XML files

To configure the OrderJoin step, you need to specify the Left list and the Right
list:
1. In the Left list, select the list whose content you need to list first. Select
Provider from the drop-down list, because you need to list the doctor
information first.
2. In the Right list, select Medication.

The OrderJoin Step is configured, as shown in Figure 9-25 on page 290.

Chapter 9. Transforming XML documents 289

 Figure 9-25 The configuration of the OrderJoin Step

The output of the OrderJoin step is a list called OrderJoin:result. This list
contains two groups: the group that you defined as the Left list and the other
group that you defined as the Right list. Therefore, as seen in Figure 9-26,
OrderJoin:result includes Provider, which contains the doctor information, and
Medication, which contains the prescribed medications. The output of the
OrderJoin step can now be mapped to a Composer step to recreate the single
XML file.

Figure 9-26 The output of the OrderJoin step

In this chapter, you configured the steps that are available in the XML Stage. We
described the Parser step in detail in Chapter 7, “Consuming XML documents”
on page 147. We described the Composer step, Regroup step, HJoin step, and
H-Pivot step in detail in Chapter 8, “Creating and composing XML documents” on
page 201.

290 InfoSphere DataStage for Enterprise XML Data Integration

 10

Chapter 10. DataStage performance and

XML
This chapter focuses on the performance aspects of processing XML documents
with InfoSphere DataStage. We begin with information about parallel processing
with DataStage, and then extend those principles to processing XML. We also
cover good design practices for maximizing performance with parallel jobs, and
documenting those patterns to avoid that might lead to poor performance and
bad resource utilization. You can obtain more information about the performance
tuning of DataStage parallel jobs in the IBM Redbooks publication, InfoSphere
DataStage Parallel Framework Standard Practices, SG24-7830, which is
referenced in “IBM Redbooks” on page 375.

© Copyright IBM Corp. 2012. All rights reserved. 291

10.1 Partitioning and collecting methods
This section provides an overview of partitioning and collecting methods, and it
provides guidelines for appropriate use in job designs. Partitioning parallelism is
key for the scalability of DataStage (DS) parallel jobs. Partitioners distribute rows
of a single link into smaller segments that can be processed independently in
parallel. Partitioners exist before any stage that is running in parallel. If the prior
stage ran sequentially, a fan-out icon is drawn on the link in the Designer canvas,
as shown in Figure 10-1.

Figure 10-1 Partitioner link marking icon

Collectors combine parallel partitions of a single link for sequential processing.

Collectors exist only before stages that run sequentially and when the previous
stage is running in parallel. They are indicated by a fan-in icon, as shown in
Figure 10-2.

Figure 10-2 Collector link marking icon

Although partitioning allows data to be distributed across multiple processes that

run in parallel, it is important that this distribution does not violate business
requirements for accurate data processing. For this reason, separate types of
partitioning are provided for the parallel job developer. Partitioning methods are
separated into two distinct classes:
򐂰 Keyless partitioning
Rows are distributed to each partition without regard to the actual data
values.
򐂰 Keyed partitioning
Rows are distributed by examining the data values for one or more key
columns, then ensuring that records with the same values in those key
columns are assigned to the same partition.

292 InfoSphere DataStage for Enterprise XML Data Integration

A special method called Auto partitioning is also the default method when new
links are created. Auto partitioning specifies that the parallel framework attempts
to select the appropriate partitioning method at run time, based on the
configuration file, datasets, and job design (stage requirements and properties).
Auto partitioning selects between keyless (same, round-robin, or entire) and
keyed (hash) partitioning methods to produce functionally correct results and, in
certain cases, to improve performance. The partitioning method is specified in
the input stage properties by using the partitioning option, as shown in
Figure 10-3.

Figure 10-3 Partitioner defined on stage input link properties

Auto partitioning is designed to allow DataStage developers to focus on the

business logic aspect of the job design and spend less time on parallel design
principles. However, the Auto partitioning method might not be the most efficient
from an overall job perspective and in certain (rare) cases can lead to wrong
results. The ability of the parallel framework to determine the appropriate
partitioning method depends on the information that is available to it. Auto
partitioning normally ensures correct results when you use built-in stages.
However, because the parallel framework has no visibility into user-specified
logic (such as within a Transformer stage), it might be necessary for the user to
specify a partitioning method manually. For example, if the logic defined in a
Transformer stage is based on a group of related records, a keyed partitioning
method must be specified to achieve the correct results.

Chapter 10. DataStage performance and XML 293

10.1.1 Keyless partitioning methods
Keyless partitioning methods distribute rows without examining the contents of
the data. The following keyless partitioning methods are available:
򐂰 Same partitioning
Same partitioning performs no partitioning to the input dataset. Instead, it
retains the partitioning from the output of the upstream stage. Same
partitioning does not move data between partitions (or, in the case of a cluster
or grid, between servers). Same partitioning is appropriate when trying to
preserve the grouping of a previous operation (for example, a parallel sort). It
is important to understand the impact of same partitioning in a specific data
flow. Because same partitioning does not redistribute existing partitions, the
degree of parallelism remains unchanged. If the upstream stage is running
sequentially, same partitioning effectively causes a downstream parallel stage
to also run sequentially. If you read a parallel dataset with same partitioning,
the downstream stage runs with the degree of parallelism used to create the
dataset, regardless of the current $APT_CONFIG_FILE.

Important: Because of its restrictive behavior, minimize the use of same

partitioning. Use it only when necessary.

򐂰 Round-robin partitioning
Round-robin partitioning evenly distributes rows across partitions in a
round-robin assignment, similar to dealing cards one at a time. Round-robin
partitioning has a fairly low overhead. Because optimal parallel processing
occurs when all partitions have the same workload, round-robin partitioning is
useful for redistributing data that is highly skewed (an unequal number of rows
in each partition).
򐂰 Random partitioning
Random partitioning evenly distributes rows across partitions. It uses a
random assignment. As a result, the order that rows are assigned to a
particular partition differs between job runs. Because the random partition
number must be calculated, random partitioning has a slightly higher
overhead than round-robin partitioning. Although in theory random
partitioning is not subject to regular data patterns that might exist in the
source data, it is rarely used in functional data flows because it has a slightly
larger overhead than round-robin partitioning.
򐂰 Entire partitioning
Entire partitioning distributes a complete copy of the entire dataset to each
partition. Entire partitioning is useful for distributing the reference data of a
lookup task, which might or might not involve the lookup stage. On clustered

294 InfoSphere DataStage for Enterprise XML Data Integration

 and grid implementations, entire partitioning might affect performance,
because the complete dataset must be distributed across the network to each
node.

10.1.2 Keyed partitioning methods

Keyed partitioning examines the data values in one or more key columns,
ensuring that records with the same values in those key columns are assigned to
the same partition. Keyed partitioning is used when business rules (for example,
remove duplicates) or stage requirements (for example, join) require processing
on groups of related records. There are a number of keyed partitioning methods
available, which are included in the following list:
򐂰 Hash partitioning
Hash partitioning assigns rows with the same values in one or more key
columns to the same partition by using an internal hashing algorithm. If the
source data values are evenly distributed in these key columns, and many
unique values exist, the resulting partitions are of relatively equal size. In the
situation where the number of unique key values is low, you can get partition
skew, where one partition receives a much larger percentage of rows than
other partitions. Skew negatively affects performance. One way of correcting
this partition skew is to add an additional key column that substantially
increases the number of unique values of the composite key, which results in
better distribution across partitions. When using hash partitioning on a
composite key (more than one key column), individual key column values
have no significance for partition assignment.
򐂰 Modulus partitioning
Modulus partitioning uses a simplified algorithm for assigning related records
based on a single integer key column. It performs a modulus operation on the
data value by using the number of partitions as the divisor. The remainder is
used to assign the value to a specific partition:
partition = MOD (key_value / number of partitions)
Like hash, the partition size of modulus partitioning is equally distributed as
long as the data values in the key column are equally distributed. Because
modulus partitioning is simpler and faster than hash, it provides a
performance improvement in situations where you have a single integer key
column. Modulus partitioning cannot be used for multiple or composite keys,
or for non-integer key columns.
򐂰 Range partitioning
As a keyed partitioning method, range partitioning assigns rows with the
same values in one or more key columns to the same partition. Given enough
unique values, range partitioning ensures a balanced workload by assigning

Chapter 10. DataStage performance and XML 295

 an equal number of rows to each partition, unlike hash and modulus
partitioning where partition skew depends on the actual data distribution. To
achieve this balanced distribution, range partitioning must read the dataset
twice: the first time to create a range map file, and the second time to partition
the data in a flow by using the range map. A range map file is specific to a
certain parallel configuration file.
The read-twice penalty of range partitioning limits its use to specific
scenarios, typically where the incoming data values and distribution are
consistent over time. In these instances, the range map file can be reused. It
is important that if the data distribution changes without recreating the range
map, partition balance is skewed, which defeats the intention of range
partitioning. Also, if new data values are processed outside of the range of a
specific range map, these rows are assigned to either the first or the last
partition, depending on the value. In another scenario to avoid, if the incoming
dataset is sequential and ordered on the key columns, range partitioning
effectively results in sequential processing.
򐂰 DB2 partitioning
The DB2/UDB Enterprise Stage (or EE Stage) matches the internal database
partitioning of the source or target DB2 Enterprise Server Edition with Data
Partitioning Facility database (previously called DB2/UDB EEE). By using the
DB2/UDB Enterprise Stage, data is read in parallel from each DB2 node. And,
by default, when writing data to a target DB2 database by using the DB2/UDB
Enterprise Stage, data is partitioned to match the internal partitioning of the
target DB2 table by using the DB2 partitioning method. DB2 partitioning can
be specified for target DB2/UDB Enterprise stages only. To maintain
partitioning on data read from a DB2/UDB Enterprise Stage, use same
partitioning on the input to downstream stages.

10.1.3 Collectors
Collectors combine parallel partitions of an input data set (single link) into a
single input stream to a stage that is running sequentially. The collector method
is defined in the stage input/partitioning properties for any stage that is running
sequentially, when the previous stage is running in parallel. The following list
shows the collector types:
򐂰 Auto collector
The auto collector first checks the dataset to see whether it is sorted. If so,
the framework automatically inserts a sort merge collector instead. If the data
is not sorted, the auto collector reads rows from all partitions in the input data
set without blocking for rows to become available. For this reason, the order of
rows in an auto collector is undefined, and the order might vary between job
runs on the same data set. Auto is the default collector method.

296 InfoSphere DataStage for Enterprise XML Data Integration

 򐂰 Round-robin collector
The round-robin collector reads rows from partitions in the input dataset by
reading input partitions in round-robin order. The round-robin collector is
generally slower than an auto collector, because it blocks each partition and
waits for a row to become available (or an end-of-data signal is received).
However, there is a special situation where the round-robin collector might be
appropriate. Assume that a job initially partitioned the data by using a
round-robin partitioner, that data is not repartitioned in the job flow, and that
the number of rows is not reduced (for example, through aggregation). Then,
a round-robin collector can be used before the final sequential output to
reconstruct a sequential output stream in the same order as the input data
stream. This approach works, because a round-robin collector reads from
partitions by using the same partition order that a round-robin partitioner
assigns rows to parallel partitions.
򐂰 Ordered collector
An ordered collector reads all rows from the first partition, blocking for data to
become available. Then, it reads all rows from the next partition, and it
continues this process until all rows in the dataset are collected. Ordered
collectors are generally only useful if the input dataset is sorted and
range-partitioned on the same key columns.
򐂰 Sort merge collector
If the input data set is sorted in parallel, the sort merge collector generates a
sequential stream of rows in a globally sorted order. The sort merge collector
requires one or more key columns to be defined, and these key columns must
be the same columns, in the same order, as used to sort the input data set in
parallel. The row order is undefined for non-key columns.

10.1.4 Choosing a collector

You use the auto collector in most situations because of its non-blocking
behavior. And, it intelligently inserts the sort merge collector when necessary.
You need to explicitly define a collector to something other than auto if you want
to force a specific collection method. Given the options for collecting data into a
sequential stream, the following guidelines form a methodology for choosing the
appropriate collector type:
򐂰 When output order does not matter, use the auto collector (the default).
򐂰 When the input data set is sorted in parallel, use the sort merge collector to
produce a single, globally sorted stream of rows.
򐂰 When the input data set is sorted in parallel and range partitioned, the
ordered collector might be more efficient.

Chapter 10. DataStage performance and XML 297

 򐂰 Use a round-robin collector to reconstruct rows in input order for
round-robin-partitioned input data sets, as long as the data set is repartitioned
or reduced.

10.2 Sorting relational data basics

Traditionally, the process of sorting data uses one primary key column and,
optionally, one or more secondary key columns to generate a sequential ordered
result set. The order of key columns determines the sequence and groupings in
the result set. Each column is specified with an ascending or descending sort
order. This method is the method with which you are probably familiar from using
SQL databases and an ORDER BY clause in a SELECT statement. When a data
set is sorted all at one time, we refer to this sort as a global sort. Figure 10-4
illustrates a global sort.

Figure 10-4 Example of a global sort operation

However, in most cases, you do not need to globally sort data to produce a single
sequence of rows. Instead, sorting is most often needed to establish order in
specified groups of data. This sort can be performed in parallel. The business
logic, such as join, can be performed accurately on each sorted group, as long as
key-based partitioning is used to ensure that all members of the group are
distributed to the same partition. Figure 10-5 on page 299 illustrates key-based
partitioning and a parallel sort.

298 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 10-5 Example of parallel sort

Notice that the partitioning happens first. The data is distributed into each
partition as it arrives. The sort is performed independently and in parallel on all
partitions. Parallel sort yields a significant performance improvement on large
data sets. Parallel sort is a high-performance solution. Perform sorting in parallel
whenever possible.

10.3 Sorting relational data for composing XML

documents
It is considerably more efficient to sort data within DataStage versus by using the
internal sort facility within the XML Stage. Relational data must be sorted in
DataStage by using the parallel Sort stage before sending it into the XML Stage
and its assembly. In multiple step assemblies, where a sort needs to take place in
the middle, it is advisable to split the design into two separate XML stages. This
way, you perform the relational sort of large amounts of data to use the
DataStage parallel sort.

When you need to compose a single XML document, the XML Stage needs to be
running in sequential mode. The partitioned and sorted relational data needs to
be collected to a single partition where the XML Stage processes it sequentially.
To collect the data to a single partition, a sort merge collector must be used to
generate a sorted sequential data set. Figure 10-6 on page 300 illustrates a job
that performs a high-performance relational sort and composes the XML
document from the result.

Chapter 10. DataStage performance and XML 299

 Figure 10-6 Sorting the relational data before composing

With the composeInvoiceXML Stage in the default configuration using the auto
collector, the parallel engine framework automatically inserts the sort merge
collector when it identifies that the upstream data is sorted. In Figure 10-6, the
sort merge collector is inserted at the point where you see the collector link
marking icon on the sortedOrders link. By choice, the user can also manually
specify a sort merge collector, as shown in Figure 10-7 on page 301.

300 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-7 Manually defining a sort merge connector

To sort the relational data, you need to specify the sort and partitioning criteria to
match the columns that we use as keys within the XML assembly. Figure 10-8 on
page 302 illustrates the configuration of the hash partitioner, which is a
key-based partitioning method. Order number is selected as the hash key, which
ensures that all the records that belong to each order are hashed to the same
partition.

Chapter 10. DataStage performance and XML 301

 Figure 10-8 Configuring the hash (a key-based) partitioning method

In the XML assembly, we need to regroup the lines within each order, and so we
sort by using order number and line number as keys. The sort keys and the
partitioning keys do not need to match exactly. The sort keys can be a super-set
of the partitioning keys; however, the order must be preserved. Therefore, the
order number must be the first sort key. Figure 10-9 illustrates the configuration
of the sort keys.

Figure 10-9 Configuring sort keys

302 InfoSphere DataStage for Enterprise XML Data Integration

Inside the XML assembly, the incoming relational data is in a flat structure, with
all of the columns in a single list. One incoming record exists for each line item,
and the order information is repeated for each row. To create the hierarchical
structure, the order elements and attributes repeat only one time per order, and
the line item elements and attributes are nested within the order and repeat for
each line item. Figure 10-10 illustrates the configuration of the Regroup Step
within the XML assembly.

Figure 10-10 Configuring the list and parent-child items for the Regroup Step

For the List to Regroup option, we select the sortedOrders list, which is the
incoming link that holds the relational data. We set the Scope option to top,
which means that the regrouped output structure shows under the top level at the
output. This choice is the only available choice for the scope property, because
the sortedOrders input list is a flat structure. The Regroup step forms a
hierarchical structure with a parent and a child list. The line items are part of the
child list, so the orderNumber is configured to be part of the parent list.

The key specification defines when an item is added to the output regroup list. In
this example, we need the regroup list to contain one item per order, so we use
the order Number as the key. Figure 10-11 on page 304 illustrates the
configuration of keys.

Chapter 10. DataStage performance and XML 303

 Figure 10-11 Configuring key and optimization for the Regroup Step

The “Input records of regroup list are clustered by key - optimize execution”
option instructs the Regroup step that the data is already sorted on the regroup
key. This sort used the parallel sort in the DataStage job prior to the XML Stage.

Important: You must select this optimize execution option to achieve the
performance benefit. If not, the XML Stage assumes that the incoming data is
unsorted and starts a less efficient, non-parallel sort.

10.4 Partitioning and parallel processing with XML

documents
XML can be processed in parallel with DataStage in two ways. The first method
is to use the partition parallelism capabilities of the DataStage Engine. For many
relatively small documents, the partition parallelism capabilities of DataStage can
help improve the overall throughput of parsing documents. While each document
is parsed in the conventional, sequential fashion, multiple documents can be
parsed simultaneously. The second method is to use the parallel parsing feature
within the XML Stage. This solution is used when a few large documents exist.

304 InfoSphere DataStage for Enterprise XML Data Integration

 One document is parsed at a time, but multiple parsers simultaneously work on a
subset of the document. The following sections discuss each method in detail.

10.4.1 Using DataStage partition parallelism

Consider a fictional coffee shop company, with more than 1,200 stores
throughout the country. Every day, as each work shift ends, sales invoice data in
the form of XML documents is uploaded from each store to the corporate
headquarters for processing. A daily DataStage job loads these documents into
the data warehouse so that figures can be available to the decision support
systems that management uses to drive its business. Each time that the job runs,
it processes many thousands of files.

To design a solution, we start with the XML files. Each file is stamped with a date
and time code, plus the store number from where it originated. This information
presents an opportunity, because many of the business calculations are driven
by store number. The DataStage job is shown in Figure 10-12.

Figure 10-12 DataStage job to process many small XML files in parallel

By using an External Source stage, we obtain a data stream of file paths to the
XML files, one path per record. In the transformer, we extract the store number
into its own column and pass it along with the full XML file path to the XML Stage.
A hash partitioner that uses the store number as the key distributes the file paths
to multiple instances of the XML Stage that run in parallel. Each instance of the
XML Stage on each partition parses an XML file for the data for one store. The
data is loaded into an IBM DB2 database table that is partitioned by store
number. Next, we describe each stage in this job.

First, the External Source stage is used, in sequential mode, to obtain a list of the
XML files in the directory to be processed, as shown in Figure 10-13 on
page 306.

Chapter 10. DataStage performance and XML 305

Figure 10-13 The external source stage can dynamically obtain a list of files

The path is defined as a job parameter by using the #parameterName# syntax.

The source program is two simple shell commands with a semicolon as the
command separator. These commands are standard for most UNIX or Linux
operating systems. They also work on Microsoft Windows, because DataStage
includes the MKS Toolkit, which provides shell programs similar to UNIX.

The Transformer stage also runs sequentially and extracts the store number, as
shown in Figure 10-14 on page 307.

306 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-14 Transformer stage extracts the store number from the file name

The transformer assembles the complete file path from the file name and the
path from the job parameter. It extracts the pure store number from the XML file
name and returns that number as a separate output column. The hash partitioner
is defined on the input link on the XML Stage by using the store number field as
the key, as shown in Figure 10-15 on page 308.

Chapter 10. DataStage performance and XML 307

Figure 10-15 The hash partitioner configured on the input link of the XML Stage

By default, the XML Stage runs in sequential mode. To configure it to run in

parallel, we open the Advanced tab on the XML Stage properties. We configured
the Execution mode property to Parallel, as shown in Figure 10-16 on page 309.

308 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-16 The XML Stage configured to run in parallel

In the XML assembly, the Parser step must be configured to accept the incoming
path from the input link and, then, open each file for parsing. This configuration is
shown in Figure 10-17 on page 310. The rest of the assembly is configured
normally, exactly as you parse a single file.

Chapter 10. DataStage performance and XML 309

 Figure 10-17 The Parser step is configured to read a file set from the input link

This solution is effective and benefits from the increased I/O throughput of
reading multiple files simultaneously. Additionally, because the XML files store
the data that is already broken down by store number, it does not need to be
repartitioned. Even in other situations where repartitioning might be required, it is
most often the case that the performance boost gained from increased
throughput more than offsets any cost for repartitioning. You can use a similar
strategy to compose XML documents, as long as each partition produces a
document or stream of multiple documents.

Important: If a single XML document needs to be created with data from

multiple partitions, that XML Stage must be run in sequential mode.

10.4.2 Parallel parsing large XML documents

In the previous section, we used DataStage partition parallelism to increase
performance with a stream of many small XML documents. If a few large XML
files exist, we need a separate approach. The XML Stage is designed to handle a
specific use case that appears frequently among extract, transform, and load
(ETL) scenarios that involve parsing large XML documents. The XML document

310 InfoSphere DataStage for Enterprise XML Data Integration

contains a single repeating list element that is a direct descendant of the root
element for the document, and there are many occurrences of this structure.
Consider a large XML document that contains invoicing information for a
business quarter. This document might contain hundreds of thousands of orders,
potentially into the millions. The document needs to be parsed, and data needs
to be extracted for each order. Figure 10-18 illustrates that the order element,
highlighted in this image in red, meets the criteria for parallel parsing.

Figure 10-18 The order list is a candidate for parallel parsing

Qualifications for parallel parsing

To enable parallel parsing, all of the following qualifications must be met:
򐂰 The XML Stage must be configured to run in parallel mode.
򐂰 The XML Stage must be configured to read the file from disk. It can be an
individual file, a set of files, or an input link that feeds it paths to a set of files.
򐂰 There must be a list (repeating element) immediately under the top level (root
node) of the document that is parsed.
򐂰 Any additional steps that follow the Parser step in the assembly must be (at
the time of writing this book) limited to the Switch step, H-Pivot step, and
Output step.
򐂰 There must be only one output link, and the list directly under the top level
must be mapped to it.

Chapter 10. DataStage performance and XML 311

 Requirement 1: Configuring the XML Stage for parallel mode
The XML Stage runs in sequential mode, by default. To switch the XML Stage to
parallel mode, edit the stage properties, click the stage icon in the image map,
and select the Advanced tab. For Execution mode, select Parallel, as shown in
Figure 10-19.

Figure 10-19 Configure XML Stage to run in parallel mode

Requirement 2: Reading the XML file from disk

The stage is not capable of parsing in parallel if the document is sent to the XML
Stage as an in-memory (or large object (LOB) reference) string. It must be able
to pick up the XML files directly from disk. The job can be designed so that the
XML Stage has zero input links, and its input files are specified directly in the
parser step. A job parameter can be used, if you want, to specify the file location
at run time. Figure 10-20 on page 313 illustrates this configuration.

312 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-20 Specify single input file

If the XML Stage is given an input link, and the records supply the document
paths, one additional configuration step is required. The partitioner on the input
link to the XML Stage must be configured to use the Entire partitioning mode.
Entire partitioning mode is necessary so that the file path is distributed to each
instance of the XML Stage that runs on each partition. Figure 10-17 on page 310
illustrates how to configure the XML Stage to read a file path from an input link.
To configure the XML Stage to use entire partitioning, edit the stage properties
and click the input link in the image map. Click the Partitioning tab, and choose
Entire from the Partition type choices, as shown in Figure 10-21 on page 314.

Chapter 10. DataStage performance and XML 313

 Figure 10-21 Configure XML Stage to use Entire partitioning

Requirement 3: Repeating element directly under the root node

It is a requirement that the repeating element must be immediately under the top
level of the document to ensure that the document can be split correctly into
multiple partitions. Figure 10-18 on page 311 shows the Output tab of the Parser
step in the XML Stage assembly. The order list, which is highlighted, is the only
structure that is a candidate for parallel parsing.

Requirement 4: Additional steps in the assembly are limited

At the time of writing this book, only the Switch step, H-Pivot step, and Output
step can be used in conjunction with parallel parsing. If any other step is used,
parallel parsing is not enabled.

Requirement 5: Single output link with mapped repeating element

The top-level repeating structure must be the only structure that is mapped to an
output link. One XML Stage can have only one partitioning method at a time.

Figure 10-22 on page 315 shows that the /invoice/order list is mapped to the
output link. If you add additional output links, or map a more deeply nested list,
such as /invoice/order/lines, parallel parsing is not enabled. Therefore, we see a

314 InfoSphere DataStage for Enterprise XML Data Integration

 problem if we want to parse in parallel but we need to access data that is more
deeply nested, such as the quantity, unitPrice, and partNumber elements from
the list of lines. In the next section, we describe a solution to this problem.

Figure 10-22 Output step mapping for parallel processing

Mechanics of parallel parsing

The parallel parsing feature works by calculating the boundaries of the repeating
element and then assigning a range of these element instances to each partition.
Each partition gets assigned a range of repeating element instances to process.
The number of partitions is defined by the number of logical nodes in the
APT_CONFIG_FILE. Each instance of the Parser step parses its own range of data
and then passes it along to the next step in the assembly.

DataStage partitioning considerations

The resulting output from the XML Stage is range partitioned, according to the
boundaries that we set during the Parsing step. Continuing further with the
invoicing example, if 100,000 orders are in the document, and a two-node
configuration file, the first node gets orders 1 - 50,000 and the second node gets
orders 50,001 - 100,000. In most cases, the downstream stage needs to
repartition the data by using a partitioning method, such as key-based hash
partitioning, to implement the rest of the business rules. Although repartitioning
overhead reduces performance, the increased throughput from the parallel
parsing almost always is substantially greater than the loss due to repartitioning.

Important: It is prudent to performance test your specific application with both

sequential parsing and parallel parsing to determine the most effective result.

Chapter 10. DataStage performance and XML 315

10.4.3 Advanced parallel parsing strategy
Continuing with the example from the previous section, a large XML document
contains the invoicing information for a complete business quarter. This
document might contain hundreds of thousands of orders, potentially millions.
We need to be able to retrieve information from all of the lines within each order.
As we explained in the rules for parallel parsing, we cannot process the line
elements in parallel, because they are not a top-level structure. Our solution is to
parse the invoice document in parallel by order, and extract each order as an
XML chunk. We can then pass the XML chunk to a second downstream XML
Stage, to parse the lines from within each order in the conventional way.
Figure 10-23 shows the initial part of the job design.

Figure 10-23 Passing a chunk of XML between stages

It is important to recognize that the second XML Stage cannot parse the lines in
parallel. The actual chunk of XML containing the order must be passed as a
string set from the first XML Stage to the second XML Stage. Parallel parsing is
only enabled when reading an XML file directly; it is not possible to parallel parse
from a string set in memory.

This design lets us get to the necessary data within the lines that we need. At the
same time, this design gives us a substantial increase in performance when
processing a large file. Right-click the order element on the Document Root tab,
and choose Chunk from the context menu. If you then click the Output tab in the
right pane, you can see that the order list element now shows that its only child
node is <>e2res:text(). The child node is a chunk of XML that is not parsed
further within this step. It is sent byte-for-byte to the next step in the assembly,
which is illustrated in Figure 10-24 on page 317.

316 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-24 Each order is passed to the next step as a chunk

The second XML Stage is set up in a typical parsing fashion. The Parser step
reads from the input link as a string set. The document root is configured to parse
the order element, as shown in Figure 10-25 on page 318.

Chapter 10. DataStage performance and XML 317

Figure 10-25 Each chunk contains one order as the document root

10.5 Tuning XML Transformation steps

Several Transformation steps in the assembly contain optimization features. To
get optimal performance, ensure that the optimization features are always used.
We describe the optimization features of the Sort and H-Join steps in this
section.

10.5.1 Sort step

The Sort step optionally can use memory for the sort. If the data that you need to
sort is so large that you cannot use this option, sort the data by using the
DataStage parallel sort and not in the XML assembly. See Figure 10-26 on
page 319, which illustrates this setting when enabled.

318 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 10-26 Enable in-memory optimization for the Sort step

If this option is not enabled, it uses disk-based files for sorting. It does not sort in
parallel. Disk-based sorts in the XML assembly must be avoided if at all possible.
The Sort step is most useful when dealing with small amounts of data that fit
completely into memory. The sort must be performed in the middle of a complex
hierarchical transformation. The sort cannot be performed by using DataStage
parallel sort in advance of or after the XML Stage.

Important: The option to use memory is absolute. After you select the
in-memory sort, you cannot fall back to disk-based sort in the event that the
data does not fit in memory. If it cannot allocate enough memory to perform
the sort, the job fails with a runtime error.

10.5.2 H-Join step

The hierarchical join (H-Join) step has an option to choose between disk-based
and in-memory-based operation. The H-Join step is not performed in parallel.
Similar to the previous statements for the Sort step, if the data does not fit into
memory, it is better to perform the join in parallel by using a DataStage join stage.
Figure 10-27 illustrates the in-memory optimization feature for the H-Join.

Chapter 10. DataStage performance and XML 319

 Figure 10-27 Enable in-memory optimization for the Join step

Important: As with the Sort step, the in-memory H-Join option is absolute.
You cannot fall back to disk-based sort in the event that the data does not fit in
memory. If it cannot allocate enough memory to perform the join, the job fails
with a runtime error.

10.6 Passing XML documents or chunks within the job

In certain situations, you need to pass chunks of XML or entire XML documents
through the DataStage job instead of reading them from files. In 10.4.3,
“Advanced parallel parsing strategy” on page 316, we described an example as
an advanced parsing technique. The following situations are additional examples
of this requirement:
򐂰 Reading or writing an XML document to or from an IBM WebSphere MQ
message queue
򐂰 Reading or writing a SOAP message body to or from a web service
򐂰 Reading or writing an XML document to or from a column in a database table

There are performance implications with these job designs. The DataStage
parallel framework is primarily designed to pass large volumes of relatively small
records. Buffering is created based on a relatively small record size. To move
data from one stage to another, the entire record must fit inside a transport block.
The default block size is 128 KB.

320 InfoSphere DataStage for Enterprise XML Data Integration

 If the record contains a column with large chunks of XML, the transport block size
likely is exceeded, resulting in the failure of the job at run time. It is possible to
increase the transport block size to accommodate larger records. It is not
possible to change the transport block size for individual links. The change only
can be made simultaneously for all the links in the job. Therefore, increasing the
transport block can result in substantially increased memory and disk space
usage. You might not experience a problem if the job has relatively few stages.
For large job designs with many stages, the job might be unusable within the
constraints of the available resources of the DataStage runtime environment. As
a solution to this problem, the Connector framework in Information Server
provides a “pass by reference” implementation for managing large character and
binary objects, which we discuss in the following sections.

10.6.1 Passing XML in-line

The default behavior as described in the previous section is that all column data
is passed “in-line” within the record structure. For small XML chunks or
documents, passing in-line is ideal when the total record size remains less than
the default block size. For small to medium job designs, increasing the block size
to be larger than the maximum document size is reasonable up to about 1 MB. To
increase the block size, set APT_DEFAULT_TRANSPORT_BLOCK_SIZE=n,
where the value of n is in kilobytes. Although it is possible, we do not advise that
you set this value higher than 1048576 (1 MB). For larger XML documents, use
the “pass by reference” feature of the DataStage Connector stages.

If the individual XML documents that are processed are all consistently close to
the maximum size, no further tuning is necessary. However, if the XML
documents vary significantly in size, with most of them less than half of the
maximum size, you might try further tuning. In this situation, set
APT_OLD_BOUNDED_LENGTH=True. This value forces both in-memory and
physical data set storage to manage all string values by their actual length and
not their maximum length. This value uses more CPU resources, but it reduces
memory and I/O.

10.6.2 Passing XML by reference as a large object column

The Information Server Connector framework can create a reference to a large
data value, and then pass this reference, rather than the actual value, within each
row. This capability allows the size of the row to remain small, so that it is
unnecessary to increase the transport block size. When this field needs to be
accessed within any downstream Connector stage, the data value is
automatically retrieved from the original source.

Chapter 10. DataStage performance and XML 321

 Important: This feature only can be used with Connector stages with LOB
column support, which includes the XML Connector.

If you attempt to access a LOB column with a non-Connector stage, it is only able
to see the reference string and not the actual content to which it refers. For
example, you cannot use the Transformer stage to modify the XML that is passed
by reference as a LOB column, because the transformer does not have LOB
column support. It sees the small reference ID string only. Appendix A,
“Supported type mappings” on page 363 contains a list of Connector stages,
which have LOB column pass by reference support, that are available at the time
of writing this book.

Important: Using references adds overhead, because the connector must

retrieve the data from the source each time that you access it.

For this reason, it is important to determine in advance the size of the XML
documents that are passed through the job, and design it to use LOB columns or
pass data in-line as appropriate:
򐂰 LOB columns with the Composer step
The Composer step supports writing data into a LOB column. In the XML
Stage assembly, open the Composer step. As shown in Figure 10-28 on
page 323, select Pass as Large Object on the XML Target properties page.

322 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-28 Output composed XML content as a LOB reference column

At the time of writing this book, the XML Connector stage uses physical disk
files stored in the scratch disk location to store XML data that is passed by
LOB reference. The downstream Connector stage, such as the DB2
Connector, retrieves the data from the file at access time.
򐂰 LOB columns with the Parser step
The Parser step supports reading data from a LOB reference column. No
special option is required, you use the String set option for XML Source in the
XML Parser Step Configuration tab, as shown in Figure 10-29 on page 324.
The XML Stage automatically detects if the column contains an in-line or
pass-by-reference value. The XML Stage takes the appropriate action to
retrieve the LOB data, if necessary.

Chapter 10. DataStage performance and XML 323

 Figure 10-29 LOB reference columns are handled by using string set XML source

Important: The Parser step cannot write a LOB reference column. All parser
output is passed by value, even when the Chunk feature is used.

If it is necessary to extract a large chunk of XML data with the Parser step, and
pass it externally to DataStage by reference, you need to add a Composer step
to manage this chunk by writing the LOB reference.

10.6.3 Managing large XML schemas with LOB references

In Chapter 7, “Consuming XML documents” on page 147, we described how to
use chunking to handle large schemas, and how we used a second Parser step
within the same assembly to process the chunk. But what if a large chunk of XML
data needs to be sent externally to the XML Stage, for example, written into a
DB2 database? Large chunks require the use of a LOB reference column;
however, the Parser step cannot write a LOB reference column. It is possible to
use the Composer step as a solution.

Step 1: Configure Parser step to create an XML chunk

The Parser step is configured to produce a chunk of XML. It is depicted in
Figure 10-30 on page 325.

324 InfoSphere DataStage for Enterprise XML Data Integration

Figure 10-30 Configure the Parser to produce a chunk of XML

Step 2: Create a generic XSD schema to hold the XML chunk

The Composer requires a document root to compose the XML. However, we
cannot expand the chunk into its elements. We want to send the chunk as a
string with the least amount of overhead possible. We can create a generic XML
Schema Definition (XSD) schema that accepts a list of any XML elements.
Figure 10-31 on page 326 contains the source for an example XSD schema that
accepts a list of any XML elements.

Chapter 10. DataStage performance and XML 325

 Figure 10-31 Generic XML schema that contains a list of any element

The <xs:any> syntax allows the schema to be extensible beyond its defined
elements, by allowing any XML element to be contained within it.

Step 3: Add a Composer step to write the LOB reference

The Composer step is configured to have the transportAny element from our
generic schema as the document root. With this design, we can map our chunk
without parsing it further, as shown in Figure 10-32.

Figure 10-32 Mapping Composer step by using a generic schema

The Composer step can then write its output as a LOB reference column, which
is accessible by any downstream Connector stage.

326 InfoSphere DataStage for Enterprise XML Data Integration

10.7 Storing XML intermediately on disk
Avoid temporarily storing or staging XML documents and large XML chunks
within persistent physical files, such as data sets. Instead, write XML directly to
its intended target for maximum performance. Data sets, file sets, and similar
stages do not have LOB column support. Therefore, XML data must be passed
in-line. In situations where you must pass XML data in line, ensure that you set
the length of the column to be the smallest possible value that can accommodate
your largest XML document. Starting with Information Server Version 8.5,
DataStage, by default, reduces I/O and storage space for physical data sets by
calculating and storing only the data in variable format instead of writing the
entire maximum length.

Important: This behavior can be disabled with the environment variable

named APT_DONT_COMPRESS_BOUNDED_FIELDS. If you currently use
this variable in your DataStage project, but you are going to write large
variable-length column values, such as XML, we highly advise that you not set
this variable at the job level.

Chapter 10. DataStage performance and XML 327

328 InfoSphere DataStage for Enterprise XML Data Integration
 11

Chapter 11. Developing XML solutions

DataStage can use multiple facilities to acquire and publish XML data. In addition
to file-based storage, XML data can be fetched from relational databases,
delivered by message queues, or consumed from Web services. Real-time and
near-real-time enterprise environments can trickle feed into DataStage jobs that
are deployed as service providers that use IBM InfoSphere Information Services
Director. This chapter describes integration scenarios that involve Web services,
message queues, and relational databases.

Many existing production deployments of DataStage use XML Pack 2.0 for
DataStage for file-based and real-time data integration. This chapter describes
several key differences of which developers familiar with XML Pack 2.0 need to
be aware when building new solutions or migrating existing solutions with XML
Transformation stage.

© Copyright IBM Corp. 2012. All rights reserved. 329

11.1 Integration with Web services
The Web Services Pack for DataStage provides stages that can read data from
and write data to external Web services with InfoSphere DataStage and
QualityStage jobs. The Web Services Pack supports the SOAP 1.1 binding over
HTTP. The Web Services Pack can internally parse and compose simple XML
structures within the SOAP header and message body of a web service request.
When the SOAP message contains a complex XML structure, such as a
sequence (a repeating element or group of elements), the Web Services Pack
stages must be used to extract the SOAP message body as an XML document,
and then use the XML Stage to parse it. The Web Services Pack also contains a
Web Service MetaData Importer. The Web Service MetaData Importer can
extract service and operation information from Web Service Definition Language
(WSDL) files and store that information in the DataStage repository. Developers
can access the information in the DataStage repository and use it to design jobs.
In the following sections, we describe an example of importing a web service
definition and parsing a complex SOAP message response by using the XML
Stage.

11.1.1 Importing WSDL and XML schema information

A WSDL document contains both the necessary endpoint information, such as
server and port number, to start the service and the metadata for the content of
both the request and the response messages. SOAP is an XML-based protocol
for Web services. The SOAP message is a well-formed XML document. To
process the SOAP message with the XML Stage, an XML Schema Definition
(XSD) schema is needed that can be imported into the schema library. There are
two possibilities. Either the schema for the message body is in a separate
schema document and the WSDL has a reference to it, or it is included in-line in
the WSDL. We explain how to obtain an XSD file for both of these conditions in
the following sections. After the XSD file is obtained, import the schema into the
Schema Library Manager. To learn more about the Schema Library Manager,
see Chapter 5, “Schema Library Manager” on page 113.

External XSD schema referenced in the WSDL

The following method is the simpler of the two situations that can be used to
obtain the XSD file. In most cases, the XSD schema can be downloaded by using
the URL in the schema location attribute from the <xsd:import> tag in the WSDL
and then imported into the schema library. Figure 11-1 on page 331 is a fragment
of a WSDL document that shows the <xsd:types> stanza.

330 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-1 External XSD schema referenced in a WSDL document

Figure 11-1 shows the section where the types are defined in the WSDL. We
highlighted the <xsd:import> tag, and it must appear as the first element in the
<xsd:schema> tag. The schemaLocation attribute of the <xsd:import> tag
identifies the location of the schema file. Normally, this location is a URL location,
and it can be downloaded with a web browser. Otherwise, it needs to be
distributed with the WSDL file.

In-line XSD schema in a WSDL document

At the time of writing this book, the Schema Library Manager reads XSD files
only, and it cannot process the WSDL directly. The schema definition needs to be
manually extracted from within the WSDL and saved as an XSD file that can be
imported. Figure 11-2 on page 332 contains an example WSDL file with an
in-line schema. This example does not include a reference to an external XSD
file. The specific contents that we need to extract are highlighted in the box.

Chapter 11. Developing XML solutions 331

 Figure 11-2 XSD schema embedded in-line in a WSDL document

Extract the highlighted <schema> element and paste it into an empty file, and
give it the .xsd extension. You must change a few lines. Add the standard XML
declaration as the first line in the file:
<?xml version="1.0" encoding="UTF-8"?>

Identify all the namespace prefixes (xsd: and impl: in this example) that are used
in the schema content. Look at the <wsdl:definitions> element at the top of the
original WSDL file to see the attributes for xmlns:xsd= and xmlns:impl=. Copy
these attributes and their assigned values into the <schema> element into the
new file. Figure 11-3 on page 333 shows the highlighted namespace prefixes
and attributes.

332 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-3 Namespace information in the WSDL definition

After you identify all the namespaces and copy everything into a new file, save
the new file with the .xsd extension. This file can now be imported into the
Schema Manager. If you receive any errors during import, the likely cause is a
missed namespace definition. Go back through these steps to verify that you
located all of the namespace prefixes and defined them in the new document.
Figure 11-4 on page 334 contains the complete XSD schema file after the
manual extraction process.

Chapter 11. Developing XML solutions 333

 Figure 11-4 Complete XSD document extracted in-line from a WSDL file

Importing the WSDL document with the Web Service MetaData

Importer
To start the Web Service MetaData Importer, open the DataStage Designer and
click Import from the main menu, expand Table Definitions, and click Web
Services WSDL Definitions. Provide the URL or file path to the WSDL
document to browse it. To import an operation, expand the plus sign (+) next to
the entry for the operation in the Web Services Explorer pane. Operations
appear with the gears icon next to them. Right-click the operation that you want
and click Import. Figure 11-5 on page 335 illustrates the Web Service MetaData
Importer.

334 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 11-5 Web Service MetaData Importer to import WSDL file

Important: You might see multiple entries for the web service in the Web
Services Explorer pane. Browse and import operations that do not end in the
suffix 12. Any operations with the 12 suffix are for SOAP 1.2, which is not
supported by the Web services stages at the time of writing this book.

11.1.2 Designing a DataStage job

After you import the XSD file for the SOAP message of the WSDL into the
Schema Library Manager, and the WSDL document is imported into the
repository, the job can be constructed. Use these high-level steps:
1. Identify the SOAP message definition and the required document root.
2. Create the DataStage job with Web Service Client and XML stages.
3. Configure the Web Service Client stage to return the entire SOAP message.
4. Configure XML parsing by using the document root identified in step 1.

Chapter 11. Developing XML solutions 335

 Step 1: Identify the SOAP message definition and the required
document root
At the time of writing this book, DataStage does not have an automatic method to
identify the SOAP message definition needed, or its document root. This step is
performed manually by examining the imported definition that is stored in the
DataStage repository. By using the DataStage Designer, browse the repository
under table definitions. Expand the WebServices folder, and expand the entry
for the web service. Locate the definition that matches the naming convention
“[Operation Name]_OUT”. Double-click to open it, and then select the Columns
tab. Look at the Description field to identify the name of the root element in the
path shown. Figure 11-6 highlights the root element.

Figure 11-6 Identify document root for the SOAP message

Step 2: Create the DataStage job with Web Service Client and
XML stages
In the DataStage Designer, create a job and add the Web Service Client stage
(WS_Client) and XML stages. Map the output of the WS_Client stage directly to
the XML Stage. Figure 11-7 illustrates the required mapping between the
WS_Client and the XML stages.

Figure 11-7 Job design fragment that shows the WS_Client and XML Stage mapping

The output link from the XML Stage can be directed to whatever downstream
stage is needed, such as a Sequential File stage or a Peek stage for testing
purpose.

336 InfoSphere DataStage for Enterprise XML Data Integration

Step 3: Configure the Web Service Client stage to return the
entire SOAP message
By default, the WS_Client stage tries to parse the SOAP message XML. To pass
it to the XML Stage, you need to configure a single output column to hold the
output message. A built-in table definition that is specifically for this purpose,
which is called SOAPbody, is in the DataStage repository. Open the WS_Stage,
select the Stage tab, and select the General tab. Use Select Web Service
Operation to browse for the operation to be used. Next, go to the Output tab,
and then, go to the Columns tab. Use Load to retrieve the SOAPbody column
definition from the DataStage repository, as shown in Figure 11-8.

Figure 11-8 Load the SOAPbody column definition

Now, click the Output Message tab. At the bottom, select User-defined
Message, and select the SOAPbody column to receive the user message, as
shown in Figure 11-9 on page 338. With this configuration, the WS_Client stage
does not attempt to parse the SOAP message body. Instead, it loads the entire
message body in-line into the SOAPbody column. Finally, complete the
configuration of the WS_Client stage as required for the web service operation.
For example, use the Input Arguments tab to define input parameter values for
the operation that is called.

Chapter 11. Developing XML solutions 337

 Figure 11-9 Configure SOAPbody to receive the entire SOAP message

Important: The Web Service Pack stages do not have large object (LOB)
column (pass by reference) support. The data in the SOAPbody column is
passed in-line as part of the record, which means that there are limits and
performance implications. For more information about passing XML data
in-line, refer to Chapter 10, “DataStage performance and XML” on page 291.

Step 4: Configure the XML Stage parsing by using the

document root from step 1
In the Assembly of the XML Stage, add a Parser step. Configure it to use String
Set as the XML Source, and select SOAPbody as the input column. For the
document root, browse to the XSD schema that was imported in the previous
section, and select the root element that was identified in step 1, as illustrated in
Figure 11-10 on page 339.

338 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-10 Configure document root for parsing SOAP message body

In the Output step, configure the mapping to obtain the list and element data that
is needed, as shown in Figure 11-11.

Figure 11-11 Map output data from the SOAP message body

Chapter 11. Developing XML solutions 339

11.2 Information Services Director: DataStage jobs as
services
When DataStage jobs are published as providers for services, you must address
a few additional job design constraints. See 3.8 “Publishing jobs as services” on
page 89 for a detailed description of designing jobs for service deployment. One
of the most important concepts is request-based processing. The submission of a
request can consist of a single row or multiple rows of data. Also, you can have a
single input row that has a column that contains XML data that is parsed into
multiple rows. In these situations, DataStage must be able to determine which
rows belong to which request. To determine which rows belong to which request,
DataStage inserts a marker in the job data stream after the last row for each
request. This marker is commonly referred to as the end-of-wave marker.

11.2.1 Understanding how end-of-wave works

Records move through the DataStage job in a first-in, first-out order. DataStage
places a marker in the stream of data, after the last row of each group of rows. By
watching for this marker, DataStage stages can act to ensure that each group of
records is treated independently, mimicking what happens if the job runs with
only those rows as input. The end-of-wave marker is not actually part of the data,
and it does not get written to any external target. However, it is visible in the Peek
stage for debugging purposes, and it is shown as the following sequence of
characters: “:-- ”. You can obtain additional information about end-of-wave
behavior for real-time jobs in 3.7.2 “End-of-wave” on page 80.

When working with XML documents inside of a job that is deployed as a service,
the end-of-wave construct is critical. If a group of rows is consumed to compose
an XML document, it must stop when the end-of-wave marker is reached. When
parsing incoming XML documents, it must output the end-of-wave marker at the
end of each input document that is consumed.

Important: DataStage parallel jobs must be configured to use only a single

partition when they are used as providers for Information Services Director.

At the time of writing this book, the DataStage parallel engine framework cannot
automatically force stages to run on a single partition when the job is deployed as
a service provider. It is the responsibility of the developer to ensure that the job is
configured to use a single-node configuration file at all times. Failure to configure
the job to use a single-node configuration file at all times results in potentially
incorrect output from the service operation.

340 InfoSphere DataStage for Enterprise XML Data Integration

11.2.2 Designing and testing DataStage jobs with end-of-wave
Test-modify development cycles can be cumbersome for a job that must be
deployed as a service. Before going into production, always use a test harness
with the same type of binding that you use in production. A simpler solution that
is effective for use during the development cycle is to use a stage called Wave
Generator, which is included with DataStage. Wave Generator is on the
Processing tab of the stage palette in the DataStage Designer. You can use this
stage in a classic non-service mode DataStage job to simulate the end-of-wave
that is produced in real-time mode. Then, you can verify in advance that the jobs
work as designed. An example is shown in Figure 11-12.

Figure 11-12 Non-Information Services Director DataStage Wave Generator testing job

You do not need to perform any specific configuration in the XML Connector
stage to handle the end-of-wave, it is handled automatically. Insert Peek stages
at various points in the job design so that you can observe the end-of-wave
marker and verify that the correct output is generated.

11.2.3 Expected Peek stage output when parsing a document

The Peek stage shows all of the relational records that are parsed from the first
document, followed by the end-of-wave marker, followed by all the records from
the next document, and so on. See Figure 11-13 on page 342 for an example of
how the output appears in the DataStage Director log.

Chapter 11. Developing XML solutions 341

 Figure 11-13 Peek results from parsing with end-of-wave

11.2.4 Expected Peek stage output when composing a document

The Peek stage shows the first entire XML document, whose contents consist of
only the relational records from the first request, followed by the end-of-wave
marker, followed by the next document, and so on. See Figure 11-14 for an
example of this output in the DataStage Director log.

Figure 11-14 Peek results from composing with end-of-wave

Important: If your XML Stage has multiple output links, the end-of-wave
marker is included in the output. The marker follows the records from each
document on each link.

11.3 Event-driven processing with IBM WebSphere MQ

In addition to Information Services Director, you can also implement real-time
and near-real-time applications by using IBM WebSphere MQ and InfoSphere
DataStage. The DataStage MQ Connector stages can browse and write to
WebSphere MQ message queues, Java Message Service (JMS) queues, and
topics. When used in conjunction with the Distributed Transaction Stage, a unit of
work (XA transaction) can be managed between multiple MQ queues or a
supported relational database. For more information about transactional
event-driven processing, see 3.8 “Publishing jobs as services” on page 89, which
discusses real-time job designs in detail.

You can use XML as the payload for MQ messages as an architecture for
event-driven applications. The MQ Connector stage supports LOB reference
columns when used both as a source and a target. This support is useful for

342 InfoSphere DataStage for Enterprise XML Data Integration

processing larger XML documents that cannot be stored in-line. Figure 11-15
illustrates an MQ Connector stage that is used as a source to browse a message
queue and retrieve XML documents.

Figure 11-15 MQ Connector stage that is configured to read XML documents

The Message options parameter is set to Yes. Set this parameter first so that it
enables other important parameters:
򐂰 Message truncation must always be set to Yes with XML documents. If this
parameter is set to No, and the message on the queue is larger than the
maximum length defined for the column, it creates multiple output records
each with part of the document. These records can cause undesirable
behavior in the rest of the job. It is better to truncate the message. Then, allow
the validation capabilities of the XML Stage to handle an incomplete
document scenario.
򐂰 Message padding must always be set to No with XML documents. This
parameter is important when LOB references are not used, and the data is
passed in-line. White space in an XML document is ignored, but these bytes
still must be stored when moving the document. Because not every message
is likely to be close to the maximum size, you waste a substantial amount of
resources by transferring irrelevant white space characters if this option is set
to Yes.
򐂰 Treat EOL as row terminator must always be set to No with XML documents.
If this parameter is incorrectly set to Yes, and XML contains any line

Chapter 11. Developing XML solutions 343

 termination characters, the MQ Connector stage produces multiple output
rows, each with a chunk of incomplete XML, instead of a single record with
the full XML document.

In Figure 11-15 on page 343, Enable payload reference parameter is set to Yes.
This value is appropriate for large XML documents. It results in a LOB reference
in the output column instead of the data. If for small documents, you want to pass
the data in-line, set Enable payload reference to No.

11.4 XML data integration with relational databases

It is a preferred practice to use the DataStage Connector stages for all database
connectivity. The Connectors have the richest capabilities for handling Unicode
data and large-sized columns, both of which are prevalent when working with
XML. Many of the major commercial database vendors include specific XML
processing functionality in recent versions. DataStage can integrate with these
features by accessing the raw data type in binary or string format, by using LOB
handling, and by using XML query functionality that is available through SQL
statements. In this document, we introduce the XML capabilities that are
available at the time of writing this book for IBM DB2 Version 9.7 and Oracle
Database 11g (11.1). We describe any specifics of which you need to be aware
when integrating these types in DataStage.

11.4.1 IBM DB2 V9 and DB2 pureXML

IBM DB2 V9 introduced extended functionality, which is called pureXML. This
function is called pureXML because DB2 can store XML in a native, hierarchical
format without breaking it down into relational components. Columns in DB2
tables that are declared with the XML data type are stored in a proprietary binary
hierarchical format. This format provides faster and more feature-rich indexing
and XML query capabilities. You can also store XML documents in a pure
character LOB (CLOB) column, which offers the fastest I/O performance if your
operations are largely limited to reading or writing the entire document.
DataStage supports both types.

The DB2 Connector can read and write to DB2 XML data type columns without
any special handling requirements. DB2 automatically converts between strings
and its internal binary storage type. If a CLOB column is used, no conversion is
required, because the XML is stored byte for byte in string format. DataStage
adds the capability to validate the XML before loading it into the database. This
capability eliminates the overhead of configuring an XML schema in the
database for validation, which impacts load performance. It also allows reject
handling to be managed in the ETL process.

344 InfoSphere DataStage for Enterprise XML Data Integration

 DB2 Connector XML query support
The DB2 Connector supports full and partial XML document queries by using
SQL and SQL with XML extensions (SQL/XML). The result can be retrieved
either in-line or by using LOB reference columns:
򐂰 Pure SQL queries can return the complete contents of the column only, as
depicted in Example 11-1.

Example 11-1 SQL query that retrieves an XML column

select xml_order from orders_table

򐂰 With XML query extensions, you can return a specific piece of XML from
within the column or reference a value of a specific element in a where clause
predicate. Example 11-2 shows this type of a query.

Example 11-2 SQL/XML query

select xmlquery('$c/Order/Customer/Address' passing xml_order as
“c”) from orders_table

Important: The DB2 Connector needs to be able to reconcile each object in

the query result with a column defined in the stage. When the developer uses
expressions, such as xmlquery, the developer must either use an 'as' clause
in the SQL query to assign a name to the expression or use the expression
prefix feature in the DB2 Connector stage.

It is also possible to use the XQuery language instead of SQL in the DB2
Connector. XQuery restricts the result to be retrieved as an in-line value; the DB2
Connector is unable to use a LOB reference column in this case.

11.4.2 Oracle 11g and Oracle XML DB

Oracle identifies its XML features by using the terminology XML DB. The features
are similar to the pureXML features for IBM DB2. Oracle 11g (11.1.0) introduced
new capabilities, such a binary storage type, and new XML indexing capabilities.
The Oracle data type XMLType allows the Oracle database to be aware that XML
is stored in the column and enables XML-specific features. With Oracle XML DB
11.1, the XMLType has three storage types: Structured (object relational),
Unstructured (CLOB), and Unstructured Binary XML. See “Related publications”
on page 375 for links to Oracle-published reference documents to learn more
about choosing the storage type that is best for your type of application.

Chapter 11. Developing XML solutions 345

 Retrieval: Regardless of the underlying storage type that you use for the
XMLType column, you can bring back the data at query time in multiple ways.

If the data is small enough to fit in the standard Oracle string buffer, you can bring
it back as varchar. Otherwise, you need to bring it back by converting it to a
CLOB or binary LOB (BLOB). Check your Oracle database documentation to see
the size of the Oracle string buffer for your specific release.

Reading XMLType data as CLOB

For data that is larger than the standard string buffer, you must retrieve it by
converting to an Oracle CLOB type. Use the getClobVal() function in the SQL
statement. DataStage can either store the CLOB in-line, or pass by reference.
See Chapter 10, “DataStage performance and XML” on page 291 for more
information about whether to use pass by reference or passing in-line. With the
getClobVal() function, you must also define table and column aliases in the
statement. Oracle requires the table alias as part of its SQL syntax. The column
alias is required by the Connector stage, so that it can reconcile the objects in the
SQL result against the objects defined on the link. Figure 11-16 illustrates how to
compose the SQL statement in the Oracle connector.

Figure 11-16 Retrieve XMLType object as a CLOB

The following examples show incorrectly formed queries. The query in

Example 11-3 is incorrect. It contains the column alias, but it is missing the table
alias.

Example 11-3 Query without table alias

select docid, document as document from order_xml;

346 InfoSphere DataStage for Enterprise XML Data Integration

The query in Example 11-4 is incorrect. It contains the table alias, but it is
missing the column alias.

Example 11-4 Query with table alias but missing column alias
select t.docid, t.document.getClobVal() from order_xml t;

The query must use both alias types, as shown in Figure 11-16 on page 346, or a
runtime error message occurs.

Using XPATH in an SQL query

If you use XPATH statements as part of the SQL query, for example, by using the
extract() function to retrieve a specific portion of a document, you must also use
both a column alias and a table alias. Example 11-4 shows an example SQL
statement that extracts the first order in the invoice XML document.

Example 11-5 Query that extracts the first order from an invoice XML document
select t.docid, t.document.extract('/invoice/order[1]') as document
from order_xml t;

Inserting or updating XMLType columns

At the time of writing this book, the DataStage Oracle Connector is limited to
inserting or updating 4,000 bytes or less directly into an XMLType column,
regardless of the storage type. This limitation is imposed by the application
programming interface (API) that DataStage uses to transfer the data to Oracle.
The examples are solutions to this problem.

Solution 1: Use the ODBC Connector

The Open Database Connectivity (ODBC) Connector uses the DataDirect Wire
Protocol ODBC driver, which is included with the software. Both LOB reference
and in-line data transfer methods are supported, and they are not limited to 4,000
bytes. The ODBC Connector is the preferred approach to write large XML
documents into Oracle.

Solution 2: Implement a staging table

If you cannot use the ODBC Connector, you can use the Oracle Connector to
implement a staging table. This solution involves creating a table in Oracle that
uses a pure CLOB column instead of an XMLType column. The Oracle
Connector is configured to stage the data by inserting the pure CLOB column
into the table first. An After SQL Statement is used to insert or update the table
in the database that contains the XMLType column. Figure 11-17 on page 348
illustrates the configuration of the Oracle Connector properties to perform a
staged insert.

Chapter 11. Developing XML solutions 347

 Figure 11-17 Oracle Connector configured for a staged insert of a large XML document

The full content of the After SQL statement property is shown in Figure 11-18.
This statement inserts the staged data into the final table that contains the
XMLType column.

Figure 11-18 Insert staged CLOB data into XMLType

The job developer can customize the SYS.XMLTYPE.CREATEXML function. It

can be customized with various arguments based the storage type. It can be
customized with various arguments based on the schema that is defined for the
column in the database.

348 InfoSphere DataStage for Enterprise XML Data Integration

11.4.3 Performance considerations
For most databases, the highest throughput is obtained by writing directly to
string or binary storage types, which do not perform any conversion or
XML-related processing. Specialty XML data types are advantageous if you
actively query, modify, or index the XML data when it is stored in the database.
For more details about planning your database architecture with XML, see
“Related publications” on page 375 for links to additional resources.

11.5 Migrating from XML Pack 2.0 for DataStage

Deployments that use XML with DataStage for multiple years and across multiple
releases can have a considerable investment in jobs that use the XML Input,
XML Output, and XML Transformer stages that are part of XML Pack 2.0. These
stages are still part of the most recent DataStage release, and are available for
the maintenance of existing jobs and the development of new jobs. In most
situations, it is preferable to develop new jobs with the new XML Stage.

We explain how to move from XML Pack 2.0 stages, provide tips, and identify
pitfalls. At the time of writing this book, the Connector Migration Tool does not
provide any automated conversion of XML stages in DataStage jobs. This
conversion is a manual process. We describe the areas in which you can see the
greatest return on your investment for refactoring existing jobs to use the new
technology.

11.5.1 XML document validation and reject functionality

The XML Input, XML Output, and XML Transformer stages from XML Pack 2.0
each offered the capability to identify an output link as a reject link. This
capability can capture the reject message and the document that was rejected by
failed validation. The functionality differs substantially with the new XML
Transformation stage. As described in Chapter 6, “Introduction to XML Stage” on
page 131, the XML Transformation stage provides much finer-grained control
over validation errors, and how the stage responds to them.

Important: With the XML Transformation stage, reject functionality is limited to

the Parser step. The previous XML Pack 2.0 provided reject functionality for
both composing and parsing documents.

In most cases, validation issues can be avoided when composing documents.

You control the data that is sent into the XML Stage, and the data is scrubbed by

Chapter 11. Developing XML solutions 349

 the ETL process. For this reason, we focus on validation during the Parser step. If
you need reject behavior when you compose an XML document, you can use the
XML Pack 2.0 stages for this purpose. You need two types of information for
reject processing for an XML document that fails validation during parsing. You
need the error message that describes the failure and the document that failed.
The XML Transformation stage can capture both pieces of information.

Important: The Parser step continues to write data ahead to the next step,
regardless of whether a failure occurred.

This behavior differs from the XML Input stage, which produced no output
(except for the reject link) when a validation failure occurred. In most situations,
the job developer does not want the partially processed, and usually incorrect,
data written to the output link. In this section, we explain how to use the Switch
step in conjunction with the Parser step to implement reject functionality that is
equivalent to the XML Input stage reject function.

To use reject functionality, schema validation must be enabled. Figure 11-19 on

page 351 illustrates schema validation that is enabled and configured to perform
a reject in most conditions. This process is highly flexible, and the job developer
controls the validation.

350 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-19 Reject function enabled for validation

In this example, we send an XML instance document that contains an X

character in a numeric field. This character causes a validation failure, because
the data type has an illegal value. After Reject is selected as an outcome, the
ParserStatus value is appended to the step output, as shown in Figure 11-20 on
page 352.

Chapter 11. Developing XML solutions 351

 Figure 11-20 Parsing status information in the step output

To prevent output data when a validation failure occurs, we add a Switch step to
after the Parser step. In the Switch Step, we add a single target, as shown in
Figure 11-21 on page 353.

352 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-21 Using Switch step to check for parser validation results

The Switch step splits the output into two lists based on the specified target
condition. The isvalid list contains the output when the document passes
successful validation. The default list contains the output when the document
fails validation. The input file name is included in each list, so that you can
identify the processed document. Figure 11-22 on page 354 illustrates the
mapping for the output of the Switch step. The output link named rej gets the
validation failure message and the file name, which is mapped from the default
list. The data output link receives valid parsed data from the isvalid list.

Chapter 11. Developing XML solutions 353

 Figure 11-22 Mapping primary output and reject links from the Switch step

This design pattern achieves the same result as the reject function in the XML
Pack 2.0. At the same time, it provides additional control over the reject
conditions during validation.

11.5.2 Design patterns for composing XML documents

In this section, we describe the major design pattern differences that DataStage
developers must understand when changing jobs that compose XML documents
over to the new XML Transformation stage. The stages in XML Pack 2.0 did not
act based on the structure of the XSD schema. Instead, they had specific
hard-coded behaviors based on the XPATH and the output mode chosen by the
user at design time. The new XML Transformation stage directly depends on the
XSD schema rather than fixed behaviors. In this section, we investigate three
examples, one for each output mode from XML Pack 2.0. We demonstrate how
the equivalent functionality is achieved with the XML Transformation stage.

First, we quickly review XML Pack 2.0. The XMLOutput stage required document
boundaries (when to finalize a document and begin a new) to be managed by a
combination of the XPATH defined in the input columns and a single control

354 InfoSphere DataStage for Enterprise XML Data Integration

named output mode. Output mode allowed one of three choices: Aggregate all
Rows, Single row, and Trigger Column. This output mode, in combination with the
XPATH for the columns, controlled the document creation.

The equivalent for these behaviors in the new XML Transformation stage is
based on the use of specific steps in the assembly and the XSD. The structure
defined in the XML schema must match the instance document that you are
trying to create exactly. In the following scenarios, we demonstrate how to
configure the correct steps in the assembly, map the correct list to the top of the
document output structure, and select the correct scope on which to apply the
step.

Aggregation step in the new XML Transformation stage: XML Pack 2.0
has an output mode named Aggregate all rows. Job developers must be
careful not to confuse this mode with the similarly named Aggregation step in
the new XML Transformation stage assembly. The Aggregation step in the
assembly performs true aggregation functions, such as sum and max, on the
data. It is not used to control the creation of output documents.

XML Output: Aggregate all rows output mode

The XMLOutput stage uses a single input link only and requires a normalized
relational pattern for the data that is sent. The hierarchical relationship needs to
be represented in a single, flat, relational data structure on that input link. The
Aggregate all rows option results in a single XML document as output, with the
multiple input rows grouped as a repeating structure in the document.
Figure 11-23 on page 356 shows how the Aggregate all rows mode is selected in
the XML Output stage.

Chapter 11. Developing XML solutions 355

 Figure 11-23 “Aggregate all rows” output mode in the XML Output stage

The following example is based on a schema for an invoice. Each invoice

contains multiple orders. Each order contains multiple lines. Figure 11-24 on
page 357 shows how this example invoice structure is built with XML Output. The
XML Output stage automatically opens and closes parent elements based on
data values. When the XML Output stage reaches the element designated as
key, it produces a new element for each value of that key column. The job
developer can control the sorting of the data and the column identified as key, but
the rest is fixed behavior that is built into the stage.

356 InfoSphere DataStage for Enterprise XML Data Integration

 Figure 11-24 XML Output stage used a key column to create a nested relationship

The new XML Transformation stage continues to support this design pattern. As
described in Chapter 10, “DataStage performance and XML” on page 291, high
performance can be obtained with relational joins taking place in the parallel job
and sending sorted, normalized data into the XML Stage. In the assembly, the
normalized input data is converted to a hierarchical relationship, and it is mapped
to the target document. Figure 11-25 demonstrates how to achieve the same
invoice example with the XML StageXML Stage.

Figure 11-25 Multiple Regroup steps used to create a nested relationship

Chapter 11. Developing XML solutions 357

 The target list document_collection identifies the context for creating an XML
document. In this example, the top-level Invoice list is mapped to the
document_collection, which results in a single document that is produced for
each set of invoice data. Also, the target schema contains the lists orders and
lines, which contain the repeating (one or more elements) structure.

In the assembly before the Composer step, multiple Regroup steps are used to
create the hierarchical relationship from the flat, normalized input link. You can
see the results of these Regroup steps in the input tree view of Figure 11-25 on
page 357, labeled Regroup:Invoice and Regroup:Lines. The Regroup steps take
multiple input rows and group them into multiple element lists. The Composer
step can then easily map these lists into the output document.

Important: The configuration of the input link remains the same for both
stages. Therefore, existing jobs can be modified easily without needing to
change the job design. Merely replace the XML Output stage with the XML
Stage on the job canvas. The rest is performed in the XML Stage assembly.

XML Output: Single row output mode

In the XML Output stage, this mode results in a new XML document that is
created for each relational input row. The XPATH controls the resulting document
structure. The following example is based on a schema for an address book. The
address book stores data hierarchically by state, city, and zip code. The address
book can have multiple states; each state can have multiple cities; each city can
have multiple zip codes; and each zip code can have multiple addresses. XPATH
for the nesting relationship in the XML Output stage looks similar to the
“Aggregate all rows” mode example, as you can see in Figure 11-26 on
page 359.

358 InfoSphere DataStage for Enterprise XML Data Integration

Figure 11-26 Configuring XML Output nesting relationship for single row mode

The XPATH looks similar, because it represents only the nesting structure of the
target XML document. It does not represent any of the context, such as
minimum/maximum occurrence. The XML Output stage used the single row
output mode to enforce a single occurrence of the XML structure in each
document and the creation of multiple output documents. The key to
implementing the equivalent functionality in the XML Transformation stage is to
use the proper XSD schema. The target schema must define its hierarchy by
using groups rather than lists. In the XSD schema, set the maxOccurs attribute to
1 for the element to define the hierarchy. Figure 11-27 on page 360 illustrates the
configuration of the Composer step in the XML assembly, for the equivalent of
single row mode.

Chapter 11. Developing XML solutions 359

Figure 11-27 Mapping to a non-repeating structure

The only list in the target structure is the document_collection list, which
identifies when a new document is created. The DSLink4 input link list is mapped
to the document_collection; thus, one document is created for each input row.
The rest of the elements are then mapped. There is no Regroup step. Even
though the output target defines a hierarchical relationship, it can be managed
through mapping because of the one-to-one relationships of all the fields.

The configuration of the input link remains the same. Jobs that use the XML
Output stage with single row mode can be modified easily to use the XML
Transformation stage without having to change the rest of the job design.

XML Output: Trigger column

In the XML Output stage, this mode results in the creation of a new document
when the value of a specific column changed. The job developer can use a
column as a flag, computing the value of the column, such as in an upstream
Transformer stage, so that it changes precisely when a new document must be
created. To demonstrate this functionality, we use the previous example from
“XML Output: Aggregate all rows output mode” on page 355. In the original
example, each invoice contains multiple orders, and each order contains multiple
lines. The job creates a single invoice document that contains multiple orders.
For this scenario, we modify the example so that it creates a separate document
for each order.

360 InfoSphere DataStage for Enterprise XML Data Integration

The configuration of the XML Output stage, as shown in Figure 11-28, results in
the creation of a new output XML document whenever the value of the
orderNumber column changes.

Figure 11-28 XML Output stage configured to create a document for each order

To replicate this functionality in the XML Transformation stage, we need to

change the mapping to the document_collection target list, so that it receives a
list of orders. We also need to modify the schema so that the definition for the
invoice element contains only a single order element. To modify the schema, set
maxOccurs=”1” as shown in Figure 11-29. The order element still contains an
unlimited number of lines (maxOccurs=”unbounded”), which is the required
behavior.

Figure 11-29 Schema fragment that shows one order in each invoice document

Chapter 11. Developing XML solutions 361

 With this modified schema, the Order element appears as a group instead of a
list in the assembly, as shown in Figure 11-30. Now, the Composer step creates
only a single order in each XML document.

Figure 11-30 Mapping to create one document per order

A Regroup step is used to group the lines into a list in each order. The regrouped
Orders are then mapped to the document_collection, which results in one output
document that is created for each order.

362 InfoSphere DataStage for Enterprise XML Data Integration

 A

Appendix A. Supported type mappings

In this appendix, we provide the following tables with the supported type
mappings:
򐂰 DataStage data type to XML data type mapping
򐂰 XML data type to DataStage data type mapping
򐂰 Connector stage large object (LOB) reference support matrix

© Copyright IBM Corp. 2012. All rights reserved. 363

A.1 DataStage data type to XML data type mapping
Table A-1 provides a listing of the DataStage data type to XML data type
mapping.

Table A-1 DataStage data type to XML data type mapping

DataStage data type XML data type

BigInt long

Unsigned BigInt unsignedLong

Binary binary

Bit boolean

Char string

Date date

Decimal decimal

Double double

Float double

Integer int

Unsigned Integer unsignedInt

LongNVarChar string

LongVarBinary binary

LongVarChar string

Nchar string

Numeric decimal

NvarChar string

Real float

SmallInt short

Unsigned SmallInt unsignedShort

Time time

Timestamp dateTime

TinyInt byte

364 InfoSphere DataStage for Enterprise XML Data Integration

 DataStage data type XML data type

Unsigned TinyInt unsignedByte

VarBinary binary

VarChar string

A.2 XML data type to DataStage data type mapping

Table A-2 provides a listing of the XML data type to DataStage data type
mapping.

Table A-2 XML data type to DataStage data type mapping

XML data type DataStage data type

duration VarChar

dateTime TimeStamp

time Time

date Date

gYear Date

gYearMonth Date

gMonth Date

gMonthDay Date

gDay Date

anyURI VarChar

ENTITY VarChar

ENTITIES VarChar

ID VarChar

IDREF VarChar

IDREFS VarChar

QName VarChar

token VarChar

Appendix A. Supported type mappings 365

 XML data type DataStage data type

language VarChar

Name VarChar

NCName VarChar

NMTOKEN VarChar

NMTOKENS VarChar

NOTATION VarChar

normalizedString VarChar

string VarChar

float Real

double Double

decimal Decimal

integer Decimal

long BigInt

int Integer

short SmallInt

byte TinyInt

positiveInteger Decimal

nonPositiveInteger Decimal

negativeInteger Decimal

nonNegativeInteger Decimal

unsignedLong Unsigned BigInt

unsignedInt Unsigned Integer

unsignedShort Unsigned SmallInt

UnsignedByte Unsigned TinyInt

hexBinary VarChar

base64Binary VarChar

boolean Bit

366 InfoSphere DataStage for Enterprise XML Data Integration

A.3 Connector stage LOB reference support matrix
Table A-3 provides a listing of the Connector stage large object (LOB)
references.

Table A-3 Connector stage LOB reference support matrix

Connector stage Accepts LOB reference Creates LOB reference

CDC Connector Yes No

DB2 Connector Yes Yes

DRS Connector Yes Yes

MQ Connector Yes Yes

Netezza Connector No No

ODBC Connector Yes Yes

Oracle Connector Yes Yes

Teradata Connector Yes Yes

XML Connector Yes Yes

Appendix A. Supported type mappings 367

368 InfoSphere DataStage for Enterprise XML Data Integration
 B

Appendix B. Transformer functions

In this appendix, we present lists of the Transformer functions, which include the
Aggregate functions and Switch functions. All functions are not available for all
data types. The functions that are presented in Table B-1 are data type specific.

Table B-1 Aggregate functions

Function Description

Average Calculates the average value in the list.

Count Counts the elements in the list. Ignores null values.

Concatenate Concatenates all strings in the list, starting with the first element.

First Selects first value in the list.

Last Selects last value in the list.

Maximum Selects the maximum value in the list. If the list contains Boolean
strings, True is greater than False.

Minimum Selects the minimum value in the list. If the list contains Boolean
strings, False is less than True.

Sum Calculates the sum of all of the values in the list.

Variance Calculates the variance value of all of the values in the list.

© Copyright IBM Corp. 2012. All rights reserved. 369

 All functions are not available for all data types. The functions that are presented
in Table B-2 are data type specific.

Table B-2 Switch functions

Function Description

isNull Returns True if the value is null.

Greater than Returns true if the value is greater than the value of the
parameter.

Greater than or equal Returns true if the value is greater than or equal to the
value of the parameter.

Less than Returns true if the value is less than the value of the
parameter.

Less than or equal Returns true if the value is less than or equal to the value
of the parameter.

Equals Returns true if the two values are the same.

Between Returns true if the value is within the specified range.

IsTrue Returns true if the value is true.

IsFalse Returns true if the value is false.

Compare Returns true if the string value is the same as the string
value of the parameter.

CompareNoCase Returns true if the string value is the same as the string
value of the parameter. Ignores the case.

Contains Returns true if the string value contains the string value of
the parameter.

ContainsCaseInsensitive Returns true if the string value contains the string value of
the parameter. Ignores the case.

IsBlank Returns true if the string is empty or null.

IsNotBlank Returns true if the string is not empty and not null.

Like Returns true if the string value matches the pattern

defined by the parameter value. Use a percent sign (%) to
define missing letters before and after the pattern.

370 InfoSphere DataStage for Enterprise XML Data Integration

Glossary

Access mode. The access mode of an IBM solidDB Open Database Connectivity (ODBC). A standard
parameter defines whether the parameter can be application programming interface for accessing
changed dynamically via an ADMIN COMMAND, data in both relational and non-relational database
and when the change takes effect. The possible management systems. Using this API, database
access modes are RO, RW, RW/Startup, and applications can access data stored in database
RW/Create. management systems on a variety of computers
even if each database management system uses a
Application Programming Interface. An interface different data storage format and programming
provided by a software product that enables interface. ODBC is based on the call level interface
programs to request services. (CLI) specification of the X/Open SQL Access
Group.
BLOB. Binary large object. A block of bytes of data
(for example, the body of a message) that has no Optimization. The capability to enable a process to
discernible meaning, but is treated as one entity that execute and perform in such a way as to maximize
cannot be interpreted. performance, minimize resource utilization, and
minimize the process execution response time
DDL. (Data Definition Language). An SQL delivered to the user.
statement that creates or modifies the structure of a
table or database, for example, CREATE TABLE, Partition. Part of a database that consists of its own
DROP TABLE, ALTER TABLE, and CREATE data, indexes, configuration files, and transaction
DATABASE. logs.

Disk-based table (D-table). A table whose contents Primary Key. Field in a table that is uniquely
are stored primarily on disk so that the server copies different for each record in the table.
only small amounts of data at a time into memory.
Process. An instance of a program running in a
DML. (Data Manipulation Language). An INSERT, computer.
UPDATE, DELETE, or SELECT SQL statement.
Server. A computer program that provides services
Distributed Application. A set of application to other computer programs (and their users) in the
programs that collectively constitute a single same or other computers. However, the computer in
application. which a server program runs is also frequently
referred to as a server.
Dynamic SQL. SQL that is interpreted during
execution of the statement. Shared nothing. A data management architecture
where nothing is shared between processes. Each
Metadata. Typically called data (or information) process has its own processor, memory, and disk
about data. It describes or defines data elements. space.

Multi-Threading. Capability that enables multiple SQL passthrough. The act of passing SQL
concurrent operations to use the same process. statements to the back end, instead of executing
statements in the front end.

© Copyright IBM Corp. 2012. All rights reserved. 371

Static SQL. SQL that has been compiled prior to
execution. Typically, it provides the best
performance.

372 InfoSphere DataStage for Enterprise XML Data Integration

Abbreviations and acronyms
IBM International Business DML Data Manipulation Language.
Machines Corporation An INSERT, UPDATE,
ITSO International Technical DELETE, or SELECT SQL
Support Organization statement.

ACID Atomicity, Consistency, DS DataStage

Isolation, Durability DSN Data Source Name
ACS access control system D-table Disk-based Table
AIX® Advanced Interactive DTD Document Type Definition
eXecutive from IBM DTS Distributed Transaction Stage
API Application programming EAI Enterprise Architecture
interface Integration
ASCII American Standard Code for EAR Enterprise Archive
Information Interchange
EJB Enterprise JavaBeans
ASP Active Server Pages
EOW End-of-wave
B2B Business-to-Business
ERE External Reference Entity
B2C Business-to-Consumer (solidDB HSB link failure
BE Back-end data server detection method)
BLOB Binary large object ETL Extract-Transform-Load
CDC Change Data Capture FE Front-end data server
CLI Call-Level Interface FP Fix Pack
CLOB Character large object FTP File Transfer Protocol
CPU Central processing unit Gb Gigabits
CRM Customer Relationship GB Gigabytes
Management GIS Geographical Information
DBA Database Administrator Systems
DBMS DataBase Management GUI Graphical user interface
System HAC solidDB High Availability
DDL Data Definition Language. An Controller
SQL statement that creates HADR High Availability Disaster
or modifies the structure of a Recovery - DB2
table or database, for
example, CREATE TABLE or HTML HyperText Markup Language
DROP TABLE. HTTP HyperText Transfer Protocol
DES Data Encryption Standard I/O Input/Output
DLL Dynamic Link Library

© Copyright IBM Corp. 2012. All rights reserved. 373

IBM International Business SBCS Single Byte Character Set
Machines Corporation SDK Software Developers Kit
ID Identifier SGML Standard Generalized Markup
IMS Information Management Language
System SL Section Leader
IS Information Server SMP Symmetric MultiProcessing
ISV Independent Software Vendor SOA Service-oriented architecture
ISD Information Services Director SOAP Simple Object Access
IT Information Technology Protocol
ITSO International Technical SQL Structured Query
Support Organization SVG Scalable Vector Graphics
JMS Java Message Services UDDI Universal Description,
JSP Java Server Pages Discovery and Integration
KB Kilobyte (1024 bytes) UOW Unit-of-Work
LDAP Lightweight Directory Access URL Uniform Resource Locator
Protocol VRML Virtual Reality Modeling
LLA Linked Library Access Language
LOB Large objects VSAM Virtual Sequential Access
Mb Megabits Method

MB Megabytes W3C Worldwide Web Consortium

MBCS Multibyte Character Set WISD WebSphere Information

Services Director
MDM Master Data Management
WSDL Web Services Description
MPP Massively Parallel Processing Language
MQ Message Queue XHTML eXtensible HyperText Markup
MVC Model-View-Controller Language
MWB Metadata WorkBench XML eXtensible Markup Language
ODBC Open DataBase Connectivity XPATH XML Path Language
OS Operating System XSL Extensible Stylesheet
Language
O/S Operating System
XSLT XSL Transformations
OSH Orchestrate SHell
PDA Personal Device Assistant
RDBMS Relational DataBase
Management System
REST Representational State
Transfer
RSS Really Simple Syndication
RTF Rich Text Format

374 InfoSphere DataStage for Enterprise XML Data Integration

Related publications

The publications listed in this section are considered particularly suitable for a
more detailed discussion of the topics covered in this book.

IBM Redbooks
The following IBM Redbooks publications provide additional information about
the topic in this document. Several publications referenced in this list might be
available in softcopy only.
򐂰 InfoSphere DataStage Parallel Framework Standard Practices, SG24-7830
򐂰 IBM InfoSphere DataStage Data Flow and Job Design, SG24-7576
򐂰 The XML Files: Development of XML/XSL Applications Using WebSphere
Studio Version 5, SG24-6586
򐂰 Implementing IBM InfoSphere Change Data Capture for DB2 z/OS V6.5,
REDP-4726
򐂰 DB2 9 pureXML Guide, SG24-7315
򐂰 Cloud Computing and the Value of zEnterprise, REDP-4763
򐂰 Master Data Management IBM InfoSphere Rapid Deployment Package,
SG24-7704

You can search for, view, download or order these documents and other
Redbooks, Redpapers, Web Docs, draft and additional materials, at the following
website:
ibm.com/redbooks

Online resources
These websites are also relevant as further information sources:
򐂰 Information Server, Version 8.7 Information Center: Oracle connector and
LOB /XML types:
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r7/topic/com.ibm
.swg.im.iis.conn.oracon.usage.doc/topics/suppt_lob_oracc.html

© Copyright IBM Corp. 2012. All rights reserved. 375

 򐂰 Oracle Database documentation: Overview of XML DB:
http://download.oracle.com/docs/cd/E11882_01/appdev.112/e23094/xdb01
int.htm#ADXDB0100
򐂰 Articles from IBM developerWorks
http://www.ibm.com/developerworks/views/data/libraryview.jsp?search_
by=infosphere+datastage+8.5
򐂰 IBM Information Center
http://publib.boulder.ibm.com/infocenter/iisinfsv/v8r5/topic/com.ibm
.swg.im.iis.ds.stages.xml.usage.doc/topics/introduction_to_xml_stage
.html
򐂰 Blog by Ernie Ostic
http://dsrealtime.wordpress.com/table-of-contents-for-reasonably-use
ful-postings/

Help from IBM

IBM Support and downloads
ibm.com/support

IBM Global Services

ibm.com/services

376 InfoSphere DataStage for Enterprise XML Data Integration

Index
Sort Merge 296–297
A Column Generator 96
ACORD 4
columns
Active Server Pages (ASP) 8
key 292, 295–298
After SQL Statement 347
LOB 322
Aggregate all rows output mode 355
mapping 171, 175, 219
aggregation, hierarchical 269
propagating 177
AlphaWorks 17
transforming values 285
Always-on stage types 79
comments 227
APT_CONFIG_FILE 76, 99, 294, 315
comments, inserting 31
APT_DONT_COMPRESS_BOUNDED_FIELDS
common connectors 74
327
common parallel processing engine 74
ASB agent 87, 90–92, 94
composeInvoiceXML stage 300
Assembly Editor 132–134, 148–150, 153, 163, 167
Composer step 322, 324, 326, 358–359, 362
Atom feeds 28
Connector stage 84, 98, 149
Auto collector 296–297
Connector stages 83, 101
Auto Map feature 176
Connectors 105
Auto partitioning 293
Constant option 178
auto-chunking 193
Constants
mapping 225
B constants, mapping 178
binary components 79 Container stages 69
binary storage 344 container stages 69
bindings, supported 13 CRM 6, 23, 103
BizTalk 6 custom plug-ins 80
block size, default 320 Customer Relation Management 6
block size, increasing 321 Customer Relationship Management 6
block size, transport 321 customer relationship management (CRM) 103
BuildOp stages 75
business intelligence xii
D
data integration capabilities 54
C data integration design 51
CDC 18 data integration issues 64
Change Data Capture 18–21 data integration logic 56
Child List option 243 data integration platform 59
chunking 193, 324 data structures 65
CLOB column 344, 347 data structures, mapping 169
Cluster topology 60 data warehousing xii
collectors database xii–xiii
Auto 296–297 database connectors 98
combining parallel partitions 292 database stage 82–84, 95–96
ordered 297 database stages 79, 82–83, 88, 95–96
Round-robin 297–298 DataDirect Wire Protocol ODBC driver 347

© Copyright IBM Corp. 2012. All rights reserved. 377

DataStage 45, 51–52, 54–56, 58–59, 65–68, 70, Extensible Markup Language (XML) 26
72–78, 85, 89–90, 98–100 Extensible Stylesheet Language (XSL) 39
components 104–106 Extensible Stylesheet Language Transformations
core capabilities 67 (XSLT) 8
jobs 59, 72–74, 89, 99 External Source stage 152
links 138, 142 extract, transform, and load (ETL) 23
parallel framework 293, 320 extract() function 347
parallel jobs 59
passing files 206
re-partitioning 72
F
File set option 152–153, 169
sharing components 72
File stages 93
stages 105–106, 108
Formatting Objects (FO) 39
utilities 110
formatting XML files 230, 248
DataStage Designer 334, 336, 341
FPML 4
DB2 149–150, 170–171
DB2 9 11–12
DB2 Connector 323, 344–345 G
DB2 partitioning 296 getClobVal() function 346
declarations 30 Grid topology 60, 63
design metadata 73 grouping stages 83
design services 74
Development/Debug stages 70
H
Distributed Transaction Stage 342 hash partitioner 301, 305, 307–308
Distributed Transaction stage (DTS) 85 hash partitioning 295, 315
document type definition (DTD) 32, 35 hierarchical aggregation 269
DTD 28–29, 32–35 hierarchical join (H-Join) 319
DTS 85 hierarchies 231
DTS stage 88, 96, 98 creating multiple levels 239
DTS stages 83, 87 creating structures 220
dynamic metadata 52, 59 HL7 4

E I
EAI 6 IBM Information Server
EJB 13, 15, 81, 85, 89, 92 execution flow 76
Electronic Business XML (ebXML) initiative 5 Server-side
Enable payload reference parameter 344 Engine 58
End-of-wave 14, 79–80, 340 Topologies supported
endpoints 89–90, 92 Cluster 62
Engine tier 54 Grid 63
enterprise application integration (EAI) 22 Three tier 61
Enterprise Architecture Integration 6 Two-tier 60
entire partitioning 313 IBM InfoSphere Change Data Capture (CDC) 18
entities 32, 44 IBM InfoSphere DataStage 1–2, 9–10, 19
environment variable 327 IBM InfoSphere Information Server 9
EOW markers 80–82, 84, 94, 96 IBM InfoSphere Information Services Director (ISD)
error messages, logging 164 55
ETL 49, 56, 64, 66 IBM InfoSphere Master Data Management (MDM)
Execution services 74 15

378 InfoSphere DataStage for Enterprise XML Data Integration

IBM InfoSphere MDM Server 15 M
IBM InfoSphere QualityStage 13 mandatory options 75
IBM InfoSphere Replication Server 20 mapping
IBM Tivoli® Workload Scheduler LoadLeveler 63 columns 171, 175, 219
IFX 4 constants 178
IMPORT statement 123–124, 128 data structures 169, 277
INCLUDE statement 125–126, 128 mapping functions 107
Information Server Connector 321 Mapping table 171
in-line passing 321 markers, inserting in a job data stream 340
input links 143–144 Marketing xii
Input stage 350 Massively Parallel Processing (MPP) 59
invalid documents 28 MDM Server 15
ISD (Information Services Director) 13 merging XML files 284, 288
supported bindings 13 Message padding parameter 343
ISD Input stage 99 Message truncation parameter 343
ISD jobs 80, 84, 87–88, 93, 95, 97–99 metadata
ISD stages 81 adding 26
ISD/JMS 88 defining for XML file 113
ISO-8859-1 character set 30 dynamic 52, 59
IXRetail 4 dynamic access 52
importing 52
management 9
J
Java Server Pages 8 managing 55
JMS 13, 15, 20, 80–81, 85–89, 92, 100, 108 operational 57
Job scores 76 project 57
Join stage 96 relational structure 150
repository 52–53, 56, 58, 60–61, 73–74,
113–114, 129
K services 50
key columns 292, 295–298 sharing 56
key-based partitioning 298, 301 storing 54
keyed partitioning 293, 295 types 49–50
keyless partitioning 294 unified 50
workbench 57
L Model-View-Controller (MVC) 8
Link stages 69 modulus partitioning 295
List to regroup option 303 MPP 59
List to Sort option 267 MQ/DTS 86–88
Large Object Columns 207 MQ/DTS (Message Queue/Distributed Transaction
LOB column 322, 327 Stage) 13
LOB reference column 323–324, 326, 345 MQConnector 80–81, 84, 86–87
LOB References 324 MQConnector stages 81
LOB references 343 multiple SQL statements 85, 98
Log once per Document action 216 multiple target types 96
Log per Occurrence action 216
logging error messages 164 N
Lookup stage 82 namespaces 25, 30, 37
native hierarchical format 344

Index 379
NewsML 4 partitioning keys 302
NMTOKENS 137–138 Pass as Large Object option 207–208, 231,
255–256
Pass as String option 206
O payload processing 14
ODBC
Peek stage 340–342
driver 347
pipeline parallelism 14, 71, 91
stage 101
plug-ins, custom 80
operational metadata 57–58
processing stages 70
Oracle CLOB type 346
project metadata 74
Ordered collector 297
pureXML 11–12, 344–345
Organizations for the Advancements of Standard In-
pureXML column 205, 256
formation Standards (OASIS) 5
OSH scripts 75
Output stage 355–361 Q
query functionality 344
P
PACKS 58–59 R
parallel framework 293 random partitioning 294
Parallel Job stages 69 range partitioning 295
parallel mode 311 real time stages 70
parallel parsing 304, 311, 314–316 recursive elements 140
parallel processing 50–51, 63, 65, 67, 70–71, 74, Redbooks website 375
76 Contact us xiv
parallel sort 298–299, 304, 319 Regroup step 360, 362
parallel sorting 299, 318 Reject action 166
Parent List option 243 relational structure, metadata 150
Parser Step 314–315, 317, 323 re-partitioning 310, 315
ParserStatus value 351 repeating element 311, 314–315
parsing Replication Server 20
parallel 311, 314–316 repository tier 54
XML files 305, 310 Representational State Transfer (REST) 78
parsing an XSLT stylesheet 198 request based processing 340
parsing schemas 185, 188 REST 78
partition parallelism 71, 304, 310 restructure stages 70
partitioning Round-robin collector 297–298
auto 293 round-robin partitioning 294
DB2 296 RSS feeds 28
entire 294, 313
hash 295, 315
key-based 298, 301
S
same partitioning 294
keyed 293, 295
Scalable Vector Graphics 7
keyles 294
schema element 37, 194, 332
modulus 295
schema elements, extracting 332
parallel 340
Schema Library Manager 113–115, 117–119,
random 294
121–123, 125–127, 129
range 295
schemaLocation attribute 331
round-robin 294
schemas 25, 28, 32, 34–35
same 294

380 InfoSphere DataStage for Enterprise XML Data Integration

 chunking 188 source 68, 70
defining 156, 194 Transformer 71, 75
error checking 128 WISD Input 92–93
generic 325 WISD Output 92–93, 95
importing 113, 259 XML 131
importing multiple 127 XML Transformer 106, 131
managing 324 stages, types 69
parsing 188 staging table 347
representation 134 Standard Generalized Markup Language (SGML)
structure 134 26
validating 159 String Set option 189, 197
XML Pack 78 Structure Validation 214
XSD 330 symbols 34
services tier 52–53 Symmetric Multiprocessing (SMP) 59
services, types 73 SYS.XMLTYPE.CREATEXML function 348
Services-based frameworks 27
Simple Object Access Protocol 17
SMP 59
T
Three-tier topology 60
SOA 53, 55, 58, 89–90
three-tier topology 62
SOAP 13, 15–17, 80–81, 85, 89, 92, 108, 320, 330,
Transformation stage 329, 349, 354–355, 357,
335–339
359–361
Sort function 267
Transformer stage 71, 75
sort keys 302
Transformer stages 75
Sort Merge collector 296–297
transforming column values 285
sorting
transportAny element 326
parallel 318
Treat EOL as row terminator parameter 343
source stages 68, 70
Two-tier topology 60
SQL statements 85, 98
Stage Editor 132
stages U
BuildOp 75 UDDI 16–17
Connector 83–84, 98, 101 United Nations Centre for Trade Facilitation and
Container 69 Electronic Business (UN/CEFACT) 5
container 69 Universal Description Discovery and Integration 17
database 82–84, 95–96 User-defined Message option 337
DataStage 105–106, 108
DTS 83, 85, 87–88, 96, 98
V
grouped 69 valid documents 28
input links 143–144 validating schemas 160
ISD 80–81 validation rules 214–215, 217
ISD Input 99 Value Validation 214
Join 96 Voice XML 6
Link 69
Lookup 82
mandatory options 75 W
MQConnector 80–81 Wave Generator 341
ODBC 101 Web 2.0 28
Parallel Job 69 Web Service Client stage 335–336
re-partitioning 72 Web Service Metadata Importer 330, 334–335

Index 381
Web Services Description Language 17 XML Input 131
Web Services Pack 330 XML Metadata Importer UI 78
WebSphere Information Analyzer 56 XML Output stage 356, 358, 360
well-formed documents 28 stages
WISD Input stage 92–93 XML Output 131
WISD Output stage 92–93, 95 XML Pack
Write to File option 208, 231, 255 for DataStage 329, 349
WS_Client stage 336–337 migrating 329
WSDL 16–17, 330–335 output mode 354–355
wsdl reject function 354
definitions element 332 schema 78
tiers 78
XML pack, 131
X XML Stage 148–156, 160, 166–167, 169–170,
XA transaction 342
188–189, 193–194, 197, 199, 265, 269, 290, 314
XBRL 4
XML Transformation stage 349–350, 354–355
XML
XML Transformer stage 131
benefits 4–5, 7
XMLOutput stage 354–355, 358
extensions 345
XMLType datatype 345
feeds 3
XPATH 347, 354, 358–359
industry standards 3, 12
XPath 39–40
native hierarchical format 344
XQuery language 345
query functionality 344
xs
stages 131, 330
list 138
syndication services 3
xsd
validation 29, 35, 160
import tag 331
XML DB 345
schema tag 331
XML files
XSD file 330–331, 335
adding comments 227
XSD schema 330
combining 276
XSLT 35, 39–41, 43
creating 202
XSLT Stylesheet, parsing 198
creating for schemas 259
creating hierarchies 231
creating in Oracle 347
extracting large chunks of data 324
filtering 265, 272
formatting 230
joining 288
metadata 113
parsing 305, 310
passing by reference 321
passing chunks 320
passing in-line 321
reading 265
recreating 290
sorting lists 267
storing 206
XML Input stage
stages

382 InfoSphere DataStage for Enterprise XML Data Integration

InfoSphere DataStage for Enterprise XML Data Integration
(0.5” spine)
0.475”<->0.875”
250 <-> 459 pages
 Back cover ®

InfoSphere DataStage
for Enterprise XML
Data Integration ®

Addresses the XML is one of the most common standards for the exchange
complexities of of information. However, organizations find challenges in INTERNATIONAL
hierarchical data how to address the complexities of dealing with hierarchical TECHNICAL
types data types, particularly as they scale to gigabytes and SUPPORT
beyond. In this IBM Redbooks publication, we discuss and ORGANIZATION
describe the new capabilities in IBM InfoSphere DataStage
Reads huge
8.5. These capabilities enable developers to more easily
documents using
manage the design and processing requirements presented
streaming by the most challenging XML sources. Developers can use BUILDING TECHNICAL
technology these capabilities to create powerful hierarchical INFORMATION BASED ON
PRACTICAL EXPERIENCE
transformations and to parse and compose XML data with
Spans both batch high performance and scalability. Spanning both batch and
and real-time run real-time run times, these capabilities can be used to solve a IBM Redbooks are developed by
times broad range of business requirements. the IBM International Technical
Support Organization. Experts
As part of the IBM InfoSphere Information Server 8.5 release, from IBM, Customers and
InfoSphere DataStage was enhanced with new hierarchical Partners from around the world
transformation capabilities called XML Stage. XML Stage create timely technical
provides native XML schema support and powerful XML information based on realistic
scenarios. Specific
transformation functionality. These capabilities are based on recommendations are provided
a unique state-of-the-art technology that allows you to parse to help you implement IT
and compose any complex XML structure from and to a solutions more effectively in
relational form, as well as to a separate hierarchical form. your environment.

For more information:

ibm.com/redbooks

SG24-7987-00 ISBN 0738436720