I'm asking a question to Datastage gurus because I'm totally lost here.
I've got a simple parallel job with an Oracle connector as a source, then a simple transformer and an ODBC stage as the target.
The target is actually an SQL Server Database.
I get wrong results in the date fields in the target compared to the source and I don't understand why.
![Embarassed :oops:](./images/smilies/icon_redface.gif)
I've tried quite a lot of things :
Trying to use a TO_CHAR function in source and then specify a varchar column in the transformer and then use a StringtoTimestamp in the transformer with the appropriate format.
Trying to disable the RCP because apparently, there is a possible mismatch in the datatypes used for dates.
And the thing is that from one execution to another, the behavior may be different. When I run the job for the same period, the results are different. Sometimes, the dates in target are wrong and just after (without touching anything but relaunching the job!), they are correct !
![Evil or Very Mad :evil:](./images/smilies/icon_evil.gif)
In the Job Logs, the only thing I got is the following :
Modified field: START_DATE, attribute: DATE_PATTERN. Design-time value: N/A. External value: N/A
Modified field: END_DATE, attribute: DATE_PATTERN. Design-time value: N/A. External value: N/A
Modified field: TRANSACTION_DATE_TIME, attribute: DATE_PATTERN. Design-time value: N/A. External value: N/A
Modified field: ORIG_TRANSACTION_DATE_TIME, attribute: DATE_PATTERN. Design-time value: N/A. External value: N/A
and all those fields are the ones that are dates and may have wrong data !
Apparently, datastage is a bit lost with the data types and is modifying something but I don't know what ! These messages are information only and not warnings...
I really don't undertsand.
If you have any clue on this, please help !
Thanks in advance,