DSXchange

08:41:56: BNA_All_Sequencer (JOB BNA_All_Sequencer) started 08:42:02: BNA_All_Sequencer finished, status=98 [Not compiled] The only documentation will tell you the same thing, 98 = not compiled. If you want to see it, look for dsapi.h in your $DSHOME/include directory. Did you do the "controll...

Exactly... I wager 400 Quatloos that it is your inserts that are the 'slow' part of this equation. You can't drive any faster than the guy in front of you.

"Date and Time formats" are fully documented in the Parallel Job Developer Guide. All is seems you are missing is the %HH for 12 hour format fixed-width hours.

Or "Unhandled abort". That and handle that fact that "BNA_All_Sequencer" is not compiled.

I'd suggest getting your official support provider involved. What 7.x version are you running? Are you certain that it supports or is supported on AIX 5.3?

No.

That's mostly because of the different jobs that would need to run each iteration: 1,2,3,4 and then 1,2,4,5. Sure it could be done but they could also easily be strung together in a 'normal' sequence, be that in a serial fashion or with some parallelism.

I think that would unnecessarily complicate things, this sounds like a simple sequence to me. Worst case you could create a new Sequence job to run the other two.

I don't see why not, just run the required jobs twice in the appropriate order (if any) for the two regions.

Find somewhere to write the file where it won't run out of room. That or write it out in segments and move each out after you write each.

And then you'll need to deal with the leading quotes in the first field and the trailing ones in the last... but some combination of the above should work.

They all have a role in that I would imagine. I have neither DataStage access nor a copy of ETLStats at the moment so can't help there. Perhaps if Kim wanders by he might be able to answer some of your questions, otherwise you'd just need to dig through the code and figure out what does what.

The FTP stages require metadata to work and thus 'specific' files. Perhaps if you generically treated each file as having one very long varchar record this could be made to mostly work as long as you set it to not add any quotes or extra characters to the file. Personally, I would prefer a scripted ...

Not what I meant. Leave the field delimiter as a pipe and tell the stage that string fields are not quoted so all quotes come in as data.