Translate Table Source
Translate table source can be changed at any time, and all rules that use that translate table reference name will get the new translate table source. The translate table source can be a z/OS file or DBMS table.
Every translate table requires format information. For z/OS files, a COBOL or PL/I layout must be identified. For translate tables stored in a DBMS, the DBMS catalog is used for format information.
Layout information is accessed when the translate table source is registered, but it is not updated if the layout changes. Therefore, if a layout is changed, the translate table layout must be reloaded.
When translate table source is changed, all disguise rules that use that reference name immediately get the new source. The fields within a translate table must be mapped to a data element to make that field available to translate actions. When a change of source is requested, validation is performed to verify that the data elements that were mapped in the original table are remapped to the new source. If all fields are not mapped, warning messages are displayed.
A long source value condition can be detected for both search source value and unique value access paths. If the value within the source field is longer than the search field in the translate table (or the unique value table entry), the long source value option applies. Choices on how to handle long source values include: source value not found (default), or truncate the value. Truncate the value drops the positions that exceed the length of the translate table field (or unique value table field) and continues with the access path as if the truncated value had been the original value. Truncated long source values are not logged or counted towards the warning threshold.
Truncating long source values may reduce the number of unique values in the target since multiple source values can now produce the same replacement value. For example, suppose source values of ABCDEF123 and ABCDEF456 were processed against a unique value table that expected these values to be six positions. Truncation would cause both values to be presented as ABCDEF and replacing both values with the same value.
Field replacement from translation always replaces the entire field. If you use a field mask, it, in effect, redefines what to process as the entire field.
The field being replaced can be a rule variable or a data element.