DSXchange

Do you know what commands you might use in such a script, or are you relying upon DSXchange to fill a gap in your knowledge of operating system commands?

This problem can be solved with a single command pipeline - you don't even need a script!

5:14pm and 8:10pm in my time zone.
Do you require that we be on line 24x7? How much would you be prepared to pay for that?!!

Write the mapping specification for each output column in English (or use FastTrack if you have it).

It's not clear from your examples how each output column is derived from the inputs.

In any case you should learn how to create/use mapping specifications.

Server jobs are much more "sensible" (less militant) about handling data like this.

Or you can use a server Sequential File stage in a server Shared Container in a parallel job to leverage the same flexibility.

Operate in sequential mode. Generate the line number. Accumulate your totals in stage variables in a Transformer stage. Generate a group number on the output, keep the line number, and use 0 as the header line number. Output header and reset totals once Mod(line_number,100) is 0. Funnel header and d...

In version 7 it might be called simply fixtool. In any case, you've proven that DS_ROUTINES is not structurally corrupted. That isolates the problem to logical corruption in one (at least) of the records in that table. Can you identify the particuar routine that throws the error?

You probably can use XMETA, but will have to work out the structure yourself. You can also get this information from DS_JOBOBJECTS table in the local project repository.

NEVER create a project in /tmp !

The location pathname must be that of a directory that does not exist, but whose parent directory does exist and to which you have rwx permission.

What "SOAP version 1.2 problem" would that be?

The first question is easily answered, yes. Use shell variables (such as $1, $2) to access the command line arguments and use these in the -param options of the dsjob command used to request the job run ( for example -param jpEndDate=$1 ). The meaning of your question about creating a table is not c...

Not out-of-the-box through to version "next" (probably 8.7).

It's easy enough to create, provided you are familiar with the "UniVerse-like" structure of DataStage projects.

DS.REINDEX only rebuilds the indices - it does not touch the underlying table. You can verify whether the table itself is corrupted using various tools, including UVFIXFILE and dsfixtool.

There's already material on DSXchange about that. Try searching,

In version 7.x the answer is No. In version 8.x the answer is Yes.

You will never get any rows output from a Sort stage until all the input rows have been consumed. This is common sense - it does not know where the final row will be in the sorted order until it has been processed.