DSXchange

If by 'once the job are running' you mean your Production environment, then yes you could get away with non-daily backups, doing it only after promotions I suppose. However, I find an automated daily backup to be more... comforting... regardless of environment. I know it will be there and I don't ha...

Well... that "comples" in the subject is a typo and should say "complete" not "compile", but I would wager you knew that.

So... in other words, if you had multiple key fields and their combined size was over the MAXKEYSIZE you would still have an issue? Dang. I knew that, honest. ps. 8) is an 8 followed by a right parenthesis, if you were wondering about Ray's "with a separation of at least 8)" statement. :wi...

64-bit ports of DataStage are few and far between. Off the top of my head, the only one I recall (having personally used it) was for Compaq Tru64 way back in the day as they had no 32-bit anything on them.

First off, it's not SQL though it looks like it is, so there's no Oracle or Toad or sqlplus for this. You need to run it in the "Universe" repository and that means either manually using the Command window of the Administrator (or from a TCL prompt) or in a job by using the UV stage set to...

Which installation, exactly - are you saying your version of DataStage is 64bit? What O/S? It doesn't matter if Oracle or your O/S is 64bit, if DataStage is 32bit you'll need to use the 32bit Oracle drivers with it.

Hashed file, not hash file.

And as Ray noted, the default max key size is 768 and I would really advise against changing that, if at all possible. If you cannot reduce the size, why not create multiple keys rather than one huge concatenated one?

Check inside your $ORACLE_HOME directory, if your version supports both 64 and 32bit libraries, two directories will be there with the default of "lib" being 64bit, especially if there is a "lib32" directory. If so, change "lib" to "lib32" in your LIBPATH stat...

You would either need to request the size of your scratch area be increased or point your job to a different, meaning larger area.

The derivations would just list the column names to be pivoted and all this should be documented in the online help available from the 'Help' button inside the stage or the pdf documentation installed on your PC.

Good, glad you were able to work that out.

Per that "standard" then, you'll never be able to use a routine that returns anything other than a zero in a Routine Activity stage without seeing a warning. You'll need to fall back on the UserVariables stage approach that priyadarshikunal mentioned, it seems, if the warning is unacceptab...

Are these thin/wired drivers that you are using? I don't see any mention of the Oracle environment variables in your dsenv LIBPATH that the thick/client-based drivers would need, hence the question.

Well, there are many types of hashed files available, and we all need to be tolerant of the... err... 'different' ones.

Still curious if the stage has the 'delete' option enabled or if something else caused the issue.

As noted, the example shows a need for a pivot, not a max() or any kind of aggregation.