DSXchange

The message about the VOC indicates that the hashed file has not been properly deleted. Again, this has been addressed before. Search the forum for DELETE.FILE command.

Have you re-booted one of the PCs to clear any scrambled DLL/OCX files from memory? What is the exact version of DataStage (use the Help menu, About topic to find out)?

DS_AUDIT does record the most recent time that an object was deleted, and by whom. Of course, if everyone logs in using the same user ID, that's not much help, though you can establish the time. LIST.DICT DS_AUDIT to get the column names for your query - from memory you need LAST.DELETED and DELETOR...

When's the interview?

Merely addressing guruji's stated uncertainty.

Since SMP environment is "share everything" pools don't really help all that much. You're right about the processing nodes. It is not actually necessary to have any pools, node or disk.

A job sequence with a "list of things" loop. Execute Command activity to initiate FTP script then Job activity to process the file.

There is a specific section in the Parallel Job Developer's Guide (page 12-6)

If you want to run them in parallel, simply emplace five job activities using five different invocation IDs in the job names. No job parameters are needed for this. The jobs can each pick up their invocation ID via the DataStage macro.

You might do well to include a Column Export stage on the reference input link to the Lookup stage, so that the 200 columns become one.

Subsequently you can parse this with the Field() function as was suggested.

Probably. Write it yourself in C, just to prove the point.

That original idea simply won't work, because DSJ.ME in a job sequence refers to the job sequence, and the job sequence "links" do not carry rows - they merely indicate a dependency. You need to create your own interlude routine that can be given the name of the job, stage and link from which to ret...

Sounds like you can mark the post as resolved, then?

PUSH the file (in binary mode) from the mainframe, rather than pulling it from the DataStage server. Use the CFF stage to read it.

You don't even need IPC to use multiple CPUs. For starters, any Transformer stage runs in a separate process from its parent job (running DSD.StageRun rather than DSD.RUN - you can see these in a listing of processes). Secondly, if you have multiple paths through the job, or even Transformer stages ...