IBM^® Rational^® Data Architect Version 6.1 - Release notes

* Although Microsoft SQL Server 2005 is not officially supported in the product, you can connect to a SQL Server 2005 database using the SQL Server 2005 JDBC Driver by following these steps:

1. In the New Connection wizard, select SQL Server 2000 from the list of database vendors.
2. In the JDBC driver field, select Other.
3. In the Class Location field, browse to the location of the Jar file for the SQL Server 2005 JDBC Driver. The default install file path is C:\Program Files\Microsoft SQL Server 2005 JDBC Driver\sqljdbc_1.0\enu\sqljdbc.jar.
4. In the JDBC driver class field, specify the com.microsoft.sqlserver.jdbc.SQLServerDriver class for the SQL Server 2005 JDBC Driver.
5. In the URL field, specify jdbc:sqlserver://YourHostName;databaseName=dbName.
6. In the User and Password fields, type your user ID and password.
7. Click Test Connection.
8. Click Finish.

** You can use Websphere Information Integrator to reverse engineer metadata from all of the relational data sources that Websphere Information Integrator supports.

Known problems, limitations, and workarounds

The following information is the currently known limitations, problems, and workarounds. The Rational Data Architect Support Web site also contains technotes and service flashes that describe changes to the documentation and known limitations and workarounds that were discovered after this document was created. The Rational Support Web site address is: www.ibm.com/software/data/integration/rda/support/

• Connecting to ODBC sources on Linux or Windows
• Compare and synchronization
• Data diagrams
• Index partitions
• Reverse engineering from a DDL file
• DDL generation
• DDL parser for DB2 UDB for z/OS
• DDL parser for Oracle
• Analyze impact
• Server discovery
• Discover function and mapping editor
• Federation support
• Delimited object names
• ClearCase
• SQL Tools

Connecting to ODBC sources on Linux or Windows

Due to a JDK problem with previous releases, you might not be able to connect to ODBC data sources using RDA on Linux or Windows unless you have DB2 Universal Database for Linux, UNIX, and Windows version 8.2 FixPak 11 or later.

Compare and synchronization

• Before you can use the data object compare editor to compare an object that is under source control, you must complete some setup steps. If you do not perform these steps, you will see an xtools error message when you try to compare objects that are under source control.
  1. Click Window > Preferences.
  2. Expand the Team folder and click Editors/Workspace.
  3. Select Use tools in the IDE if they support the selected file type in the When comparing and merging files area.
• When you are using the compare editor to merge data objects, do not edit the objects that you are comparing outside of the compare editor. For example, do not make changes in the Properties view. If you edit the objects outside of the compare editor, you might see an error message or unexpected behavior.
• After a synchronization, there are times while in the Structural View that the changes in options do not correctly refresh.
• Some results of compare and synchronization show no differences between some data type changes.
• When you compare an object in the Database Explorer with another object, and then synchronize the information, the DDL that is generated is not always accurate.
• When you are comparing objects in the compare editor and you perform a Copy from Right to Left action, the Generate Left DDL button is not enabled.
• In the following scenario, the delta DDL statement from the compare is generated incorrectly:
  1. Reverse engineer a schema that contains an alias into a physical data model in a data design project.
  2. Modify the existing alias to point to a different table.
  3. Compare the modified alias with the original (right-click the alias object, then select Compare With > Original Source).
  4. In the compare editor, click Copy from Left to Right to copy the changes to the server version of the alias.
  5. Click Generate Right Delta DDL, and save the DDL statement.
  6. The DDL statement is incorrectly generated. It includes a CREATE TABLE statement for the new aliased table even though that table already exists.

  To work around this problem, modify the DDL statement in the SQL editor before you deploy it.

Data diagrams

• The Delete from model action on a diagram shortcut object does not delete the diagram object from the model. To work around this issue, you can delete the diagram object from the model by using the Delete action in the Data Project Explorer.
• There are some limitations when saving large diagrams to an image file. Sometimes when you save a large diagram as an image file (right-click in a blank area of a diagram, and select File > Save As Image File) the image file is not created. Entries are created in the log file if logging is enabled, but there is no error message. To work around this problem, you can break up the large diagram into smaller diagrams.
• On Linux operating systems, diagram print functions do not work.

Index partitions

For zSeries Version 8 Compatible Mode database only: The partition option (Use Partition) is not supported for index partitions. For index definitions, the Storage Group option is always used.

Reverse engineering from a DDL file

For DB2 Universal Database for Linux, UNIX, and Windows and zSeries only: By default, the parser assumes that the terminator is the semicolon (";"). If the file uses a different terminator character, you must include the following statement in the first line of your DDL file:

-- <ScriptOptions statementTerminator="@" />

Where "@" is the terminator character that your DDL file uses.

DDL generation

• When you run Generate DDL for a schema with nicknames, you might see a message in the Data Output view similar to the following: Table xxx already has a primary key. The message shown in the Data Output view can be ignored.
• DDL statements for stored procedures whose name requires a delimiter in the DDL statement (for example, CREATE PROCEDURE "a.b") is not generated correctly. The delimiting quotation marks are not generated. In the example statement, the DDL is generated as CREATE PROCEDURE a.b. To work around this issue, modify the generated DDL statement to include the delimiting characters.

DDL parser for DB2 UDB for z/OS

The following DDL statements have limitations:

Statement	Limitations
ALTER TABLE	The following alterations are not supported: ADD PARTITION ADD/DROP RESTRICT ON DROP DROP MATERIALIZED QUERY
SET CURRENT SQLID	Only supported: SET CURRENT SQLID = string-constant
SET SCHEMA	Only supported: SET SCHEMA = schema-name, SET SCHEMA = string-constant

DDL parser for Oracle

• The REPLACE clause is not supported.
• TIMESTAMP is not supported as a data type when reverse engineering from a DDL file.

Analyze impact

In some scenarios when you are performing an impact analysis the product can shut down. This happens when you try to move or minimize the progress dialog or the progress bar while the impact analysis is occurring. To avoid this problem, do not move or minimize the progress dialog or the progress bar during impact analysis, and make sure to save all of your work before you perform an impact analysis.

Server discovery

On a Linux operating system, the Undefined Remote Servers do not appear for ODBC data sources, unless you create an ODBC wrapper with the wrapper name ODBC outside of the Rational Data Architect product, such as the DB2 Universal Database Control Center, or a command line. You must name the wrapper ODBC so that it is properly discovered. The wrapper on a Linux operating system is defined with a MODULE wrapper option, as in the following example:

CREATE WRAPPER odbc LIBRARY 'libdb2rcodbc.so' OPTIONS (MODULE '/usr/lib/odbc.so')

In this example, MODULE '/usr/lib/odbc.so' is the full path to the library that contains the ODBC Driver Manager.

Discover function and mapping editor

• When you launch the Discover Relationships function, be aware that aliases are treated as tables. You should decide whether to include them in the set of source schemas, or the target schema when you define the scope of the discover function.
• For the algorithms that include data sampling, only the data in Oracle and DB2 databases are sampled. To cache the sampled data, you must specify a cache database. Only DB2 UDB for Linux, UNIX, and Windows is supported as a cache database.
• On Linux operating systems, the thesaurus option for the semantic name algorithm is not supported.
• User Defined Types (UDTs) are not sampled when you discover relationships.
• In the Mapping Editor preferences, when you set the preferences for discovering relationships, the Algorithms page contains a selection for how to order multiple algorithms. You can specify the Composition by sequence or Composition by weight. When you select Composition by weight, this assigns a weighting value to each algorithm. Currently the option only applies to algorithms that return a single value.
• The SQL/XML query generation ignores the actual value of "x" of an XSD attribute maxOccurs="x"' if x is a number greater or equal to 1. The generated query will create XML elements for all rows from a source column. It will not limit the amount of selected rows to the number defined under maxOccurs. This is due to the incapability of SQL2003-conforming SQL/XML queries to express this requirement. For elements that are defined with the attribute maxOccurs="0", the mapping editor prohibits a mapping. Therefore, elements defined as maxOccurs="0" will not appear in the result.
• When UDTs are present on the target side, the generated scripts might not parse due to null value handling for UDTs.
• In the following two scenarios, not all of the artifacts are generated in the DDL script, and the script cannot be deployed without modification:
  • You are mapping from source table T1 in A.dbm to target table T2 in B.dbm, and neither A nor B are federated to an Information Integrator server. In this case, the only deployment platform available will be A.dbm and only an insert script is generated. No table object T2 for A.dbm is generated, even though this is necessary for the script to run. If you want to run the script, you must create the table.
  • You are mapping from source table T1 in A.dbm to target table T2 in B.dbm, and both A.dbm and B.dbm are federated to an Information Integrator server. In this case, the Information Integrator sever is available as a deployment option. However, if you select the Information Integrator server, only the nicknames for T1 and the insert script is generated. You must generate the nicknames for T2 from B.dbm onto the Information Integrator server before the script will run properly.

Federation support

• You can generate DDL scripts for the federated server from the Database Explorer . After you generate a script, you can deploy to like servers on DB2 Universal Database for Linux, UNIX, and Windows, DB2 Universal Database for iSeries, Oracle, SQL Server, Teradata, web services, XML, and Sybase. To deploy the DDL scripts on any other data source, you must deploy them using the DB2 command line (run them as DB2 scripts). When you deploy, you might get a message saying that the wrapper already exists. If you see this error, then disconnect the database connection and reconnect.

Delimited object names

Do not create an object that has quotation marks in the name. An object name delimited with quotation marks does not work. The following examples are not currently supported:

"""PROCEDURE"""
"""TABLE"""
"""SCHEMA"""."""PROCEDURE"""

ClearCase

• If you attempt to use the ClearCase SCM Adapter in an environment where the adapter has not been set up, the product will shut down.
• You can not expand the folders of objects in a development project that were checked out by using the ClearCase resource manager.
• You might see the following warning after you install the ClearCase Remote Client for Eclipse:

  The current configuration contains errors and this operation can have unpredictable results.

  You can ignore this warning.

SQL Tools

• The SQL Editor does not currently support host variables during the Run SQL action. To work around this issue, you can run the SQL from the SQL builder, if it is a DML statement.
• If you modify a statement in the SQL source area of the SQL builder and then you save the statement while it is invalid, the current text is not saved. Instead, the text that was in the SQL source area before modifications were made to the SQL source area is saved. If you attempt to run the invalid statement from the SQL builder, the last valid statement is run instead.
• In the SQL builder, the product does not draw lines in the graphical tables for conditions that are specified in the WHERE clause that represent a join.
• The full SQL syntax is not supported. For example, User Defined Types (UDTs) and Table functions are not supported.

Documentation Updates