We are in the process of upgrading from Cognos 10.1.1 to 10.2.2FP3, and have a large number of reports in test now failing with this error: ORA-12704: character set mismatch (error screenshot: "character set mismatch error.JPG"). The issue exists in 10.2.2FP4 as well.
This seems to be directly related to how Cognos handles:
1) NVARCHAR2 data type fields and string literals ('what you type in quotes'). I was unable to replicate this in the Samples or Audit packages until I created my own NVARCHAR2 table with 2 empty columns and was then able to replicate. Please refer to attachment "character set mismatch doc.docx" for a use case and their results in different situations.
2) Cognos macros and string literals (see screenshot: "results.JPG") Basic IF() logic seems to be failing. The full IF statement is not given in the screenshot, but then all are based on this: if(<exp> = '') then ('true') else ('false')
Has anyone else encountered this? Any recommendations? How are others working around it? We have logged the case with IBM and have APAR 75674.
Thanks,
Ed
We have also tried updating cogdmor.ini file as discussed in this thread:
https://www.ibm.com/developerworks/community/forums/html/topic?id=77777777-0000-0000-0000-000014199485 (https://www.ibm.com/developerworks/community/forums/html/topic?id=77777777-0000-0000-0000-000014199485)
The cogdmor.ini changes did not assist with the character set mismatch issue.
Only 1 attachment allowed per post...
I also started a thread for this issue here: https://www.ibm.com/developerworks/community/forums/html/topic?id=a795d9f4-1c82-4f37-bca6-3f484cea61b9#repliesPg=0