DSXchange

Then you can't use a file because of concurrency issues. If you can't use a DB sequence, then look into the Server SDK KeyMgnt routines to generate a Surrogate Key, in particular the 'concurrent' version.

Does each job need to maintain a unique sequence for itself, or do you need one sequence that is used across all jobs?

Or that.

Well... technically it is a command. See man uname for the particulars.

Since this is basically a QualityStage question, moved it here.

No really. It must be stored somewhere, incremented and then passed to (or fetched by) any job that needs it. Not sure that the nuances of each step really change that basic fact.

qt_ky wrote:How about in a Parallel job using an inverse pair of Convert() functions--one for each output column?

Sure seems like that would work!

I doubt that they always occur in the same spot each time, hence the question on how to detect them.

Exactly... you'll also have to ensure that they are multi-instance capable, meaning designed to be able to run in that fashion. As noted, that means more than just checking a check-box, although for simple designs that can be all it takes. It's all about shared resources, targets, lookups (etc) you ...

jonesv wrote:we may have stress testing going on concurrently with function testing.

At first blush that doesn't make much sense to me... 'stress' testing at the same time that other tests are underway?

Off we go to General / Sequence job land...

Then it's up to you if you want to go to all the trouble of reading them all as one long string and then parsing each one "in job" once you determine what structure it is.

Not what I said... do they all share the same structure? Same format for headers, detail, etc with the only difference being the name of the file?

Output where? To what datatype? You'll need to format it as a string to remove the natural decimal attributes you are showing now.

As long as they share the same structure / metadata, use the File Pattern option.