The source code page depends on the type of source:
Flat files and VSAM files. The code page of the data in the file. When you configure the flat file or COBOL source definition, choose a code page that matches the code page of the data in the file.
XML files. The PowerCenter Integration Service converts XML to Unicode when it parses an XML source. When you create an XML source definition, the PowerCenter Designer assigns a default code page. You cannot change the code page.
Relational databases. The code page of the database client. When you configure the relational connection in the PowerCenter Workflow Manager, choose a code page that is compatible with the code page of the database client. If you set a database environment variable to specify the language for the database, ensure the code page for the connection is compatible with the language set for the variable. For example, if you set the NLS_LANG environment variable for an Oracle database, ensure that the code page of the Oracle connection is identical to the value set in the NLS_LANG variable. If you do not use compatible code pages, sessions may hang, data may become inconsistent, or you might receive a database error, such as:
ORA-00911: Invalid character specified.
Regardless of the type of source, the source code page must be a subset of the code page of transformations and targets that receive data from the source. The source code page does not need to be a subset of transformations or targets that do not receive data from the source.
Select IBM EBCDIC as the source database connection code page only if you access EBCDIC data, such as data from a mainframe extract file.