DSXchange

You'd have to explain what your pivoting rules are in plain English. Right now, the output example is confusing. Did you leave out a row that should show "3,4,5"?

You can't leverage any "row data" in the pre or post areas, which is basically what you are trying to do. You'd either need to hard-code any values there or perhaps use job parameters to control what values get deleted.

So what kind of "few records" do you need to delete?

Specifically: viewtopic.php?t=149088

DataStage is not a Universal Translator. While it can convert from one characterset to another, it cannot translate from one language to another, which is in no way the same thing.

Some random thoughts...

I have no idea what the maximum size of the "list of things" can be for the Start Loop stage to iterate through.

Do you really want to be running up to 100 PX jobs at the same time?

That methodology won't allow you to monitor the outcome of the job runs.

"Parse them to run in a loop" means what? Are we talking about a Sequence job with the Start and End Loop stages? Or are we talking about a fully scripted command line solution? Since you mentioned the Execute Command stage I'm assuming the former but not sure what you mean by "dsrunj...

Jayanto wrote:Thanks.. But what function to use??

DecimalToString. And if you still need to, Convert after that.

Output it as a string so you have full control over the formatting.

@DATE returns the internal value for the current date.

Right, you have to use DSAttachJob first to establish a handle, then pass that handle to the other functions. And lastly detach it when you are done.

A Universe table? No, there's one per job. If you are on 8.7+ there is the Operations Console or you could script something using dsjob from the command line.

"Before job" actually runs after the job starts but before anything on the canvas runs, so you should be able to find the pids there.

Database? No. I'm not sure if the "instant" client works or not, perhaps others do. Full client works just fine, of course.

You'll need the client installed on the ETL server for any native connectivity. Otherwise you'll need to go ODBC, say with the Oracle thin/wire drivers which don't require a client.

Download the IBM Redbook pdf on InfoSphere DataStage Parallel Framework Standard Practices, lots of great info there and a whole chapter on partitioning.