Handling Multibyte Characters
When the Oracle database is created, the database charset is set to UTF-8. In addition to the UTF-8 charset, the TIBCO MDM database needs to be deployed with a configuration for character semantics that enable globalization and enhance code portability.
From Oracle9i onwards, you can set the default character semantics at either the session or instance level using the NLS_LENGTH_SEMANTICS parameter. This parameter must be effective before a table is created.
The NLS_LENGTH_SEMANTICS parameter decides how strings should be stored in the database, as characters or bytes. TIBCO recommends that strings in the database be stored as characters (CHAR). This setting takes care of globalization issues with strings.
- For a New TIBCO MDM Installation
The database object creation scripts specify the correct semantics. If you use these scripts, you do not have to perform any additional steps.
- For an Existing TIBCO MDM Installation
Prior to TIBCO MDM 7.1, TIBCO MDM instances were created using the seed data export dump provided. Depending on the database NLS_LENGTH_SEMANTICS, the tables may not have been created with the correct semantics to handle NLS characters. You can convert the older instances to the correct semantics as follows:
To convert an existing schema from byte semantics to character semantics and a single-byte character set to a multibyte character set: