DSXchange

Can you see/open the particular job in Designer, Manager, Director? If so, you should be able to delete it (once closed) from the GUI. Otherwise, you will need to delete references to it from DS_JOBS and DS_JOBOBJECTS using SQL, and to remove the repository tables using recursive rm commands. Make s...

Duzn't matta!

Parallel jobs on Windows actually run on MKS Toolkit (a UNIX emulator). So you have all the UNIX shell functionality you need. The tr command is definitely available.

Can you wrap these designs in Code tags, so that you can get the formatting right (and therefore comprehensible) please?

You can do it in one job. Perform the full outer join, then use a downstream Filter stage to constrain the rows passed through. I needed to be certain of what you needed to do.

The API is there so you can CODE lots of extra stuff. You still need to compile within the DataStage tool, so it can record the information it needs about what you've done. And, if you find my answers unacceptable, please don't let me delay your search for someone who cares about that particular fact.

Are NLS_LANG and LANG_C environment variables correctly set?

I'd use a longer VarChar for the ID column; possibly VarChar(254) to be on the safe side.

Maybe they never included DataStage in the auto-start-on-reboot directories (/etc/rc2.d for example)? Or maybe 5.9's system libraries are differently located than 5.8's. Involve your DataStage support provider for specific advice relating to your site.

Welcome aboard. :D

Perform a lookup against the target table, or a hashed file containing relevant rows and columns from the target table. This will allow you to determine the current values, which you can use in the UPDATE statement if the source value is null.

Each job type in 7.5 will be whatever it was in 5.2.

Ad hoc or predictable updates?

Only by having a master master sequence that does not abort under any conditions.

Connection refused can be other things, such as port 31538 not being allowed through a firewall. You need to undertake a systematic investigation of where the problem is. For example, is the server machine running? Is dsrpcd running? Is port 31538 recorded in /etc/services? And so on.

Almost certainly, but there's not enough information in your post to suggest better mechanisms. Stage variables are evaluated for every row processed, so it's not surprising that there are repeats - a null in a new row in the same column will generate a repeat (if I read your specification correctly).

Semi-colon should work. It's the fundamental command separator in every UNIX shell of which I am aware. Please post your exact syntax.