DSXchange

No, you don't need a "signon" file to make this work. While you could certainly use one, it is not required. You will need to get your syntax correct, however, and the post you linked to had good information on the subject in addition to what's already here. Did you try adding the "-d...

Either of those approaches can be made to work, but why run a job in a before-job subroutine? That's what Sequence jobs are for.

Please be mindful of the forum you post in. The FAQ Discussion forum was not appropriate for this post, so I moved it here. Please edit your post and add your particulars - version, O/S, etc.

If you know the maximum possible number of values, it's a horizontal pivot of rows to columns using the Pivot stage. First break that one field out into multiples, pivot and then remove any null values from the result.

Ha - got distracted watching Ani-Monday.

Job names are unique, it knows. Use the syntax others have posted, don't make stuff up. And lastly... yes. Provided you get everything else right, of course.

dsjob uses the project name and job name, do not supply any category information. And without one of the "wait" options it will just start the job and exit back to the command line.

:? If you say so. Iconv converts the date into an internal format, which is an integer value representing the number of days past the zero date. You really need OConv to get it back to the valid external format that your post shows - unless your target is a database column. Some will accept that int...

Something like this:

If we're talking Oracle here, spaces get trimmed and you still end up with a null in the target. For that, add $APT_ORACLE_PRESERVE_BLANK to your job and set it to True.

Why not make sure you are current on patches, get that possibility off the table? IBM will want you to have done that first before you go to them for any help.

Or just do it like you have posted:

That log is corrupted, most likely from it having "too many" records in it. An exact search for "computed blink of" will turn up many conversations on the topic and the solutions you have.

And? Three questions, two answers, one big hint.

How many nodes? How are you partitioning in your job? Is it consistent if you run on one node?