DSXchange

I too would suggest the 'zip' approach to this, who wants to get an email with 95 files attached to it?

Already noted.

For me, the posts were in the morning and afternoon of my first day working for a new client. Which means if I'm here, it will be either first thing in the morning or perhaps late at night (or both) but not during my working hours.

Then best to involve your official support provider, it would seem. Curious, is this a problem for everyone or just you?

Thanks for the information Scott.

This is the data "masking" they've talked about at IOD, yes? If so, I understood that it is as you said, only if you engage them will they allow access to the functions they have written. Did seem very nice, though.

And I do think they may have echoed the 'someday' comment.

No, best to ask your official support provider for specifics on that. I assume you could just reapply it, but it may not be necessary.

ps. There is a $Counter variable available that the Start Loop stage automatically manages, it may help you derive unique filenames each loop.

Did you install NLS when you installed DataStage? If not, you'll need to reinstall DataStage AFAIK, otherwise I'm sure the process to 'enable' it is well documented.

Just create another job parameter and use it in the sequential filename box surrounded by hash/pound signs.

If your "source target date" comes in as a full timestamp, you'll need to adjust your transform code slightly to chop off the time and just work with the date portion: Target Date = Oconv((Iconv(TaskInfo.SO_TARGET_DATE[1,10], "D-YMD") +1), "D-YMD[4,2,2]") : " 00:00...

"in job suppose i am getting error i need to compile the job again"

No, no you don't. You Reset aborted jobs, not recompile them.

Perhaps your characters are not "unprintable"? Are they actually a single control character or are they two characters, a "^" and a "C" for example like you typed in?

Then simply parameterize the name of the output file as well and change it appropriately each iteration of the loop.

Well, then you'd have to "decode" it back. As noted, declare the timestamp as a string, the format it is coming in is directly collatable so will aggregate just fine.