NLS with DS 3.63r2

[admin]

We are having an issue with NLS when connecting to Unidata databases that contains accented characters. The environment is as follows: DataStage 3.63r2 running on HP-UX version 11.0 on a N4000 machine. All Unidata databases are running release 4.1.36 (US, France, Germany, etc.) (Have not upgraded to DS 4.0 or higher because we were told that DS 4.0 does not support NLS with Language Group 2 and 3.63 supposedly does support NLS with Language Group 2)

The issue surrounds special accented characters that are used in French, German, etc. When we view the data from within Unidata, we see the accented characters. However, when we view the data from within DataStage utilizing the UDT stage, we see the special characters replaced with question marks. When we execute the jobs, they typically abort when the special characters are located at the end of the line but this is not always consistent. We have seen in a couple of examples that if there are two question marks (meaning that the tool found two special characters) then it tries to read the next attribute and aborts at that point. We have tried to push data out to a text file and to a hash file with no luck.

Every time the job aborts, the following error message is found in the DataStage log:
"UDTGetNext() NLS mapping error".

We have performed the following:
1. Added additional NLS mappings/locales through DSAdministrator 2. Stopped and restarted DataStage services to load new mappings. 3. Reimported the metadata in DSManager for the UDT stages and specified the new NLS mappings. 4. Rebuilt the jobs and reimported the new metadata. We also used different mappings in the NLS tab of the UDT stage (doesnt seem to matter if we use FR-French or US-English as a locale).

I searched Informix/TechInfo and found nothing that applies to this issue. The online manuals for DataStage do not go into specific troubleshooting issues for NLS.

Any information to help on this issue would be greatly appreciated. This is a major stumbling block - we have performed the loads of the US data and are now ready to move on to other countries.

Thanks
Stacy Scoggins