DSXchange

When you get version 8 there are inbuilt performance analysis and resource estimation tools.

I'm guessing that the master record is the level 01 item.

Help us to help you. What is jpfile? What is your job design? What are the permissions to jpfile and its parent directories with respect to the executing user ID?

That information should come on documentation from the vendor from whom DataStage was purchased, ultimately from IBM. In exchange, I might add, for money.

Rubbish. At the very least the score will be in the log. This is one of your primary diagnostic tools - it's what is actually executed.

Point it to the correct version (8i, 9g or 10g) in the lib directory.

Resist stupid requirements!

Your professional opinion is that the optimal approach is to separate inserts from updates.

My guess is that you are attempting to use a routine that includes BCI functions and that your 30 day grace period on the licence has expired. You must now purchase a licence from Data Direct for this combination to be able to work again.

An interactive debugger for parallel jobs is in "the plan", probably for the release after next. Caveat: that's not a promise that one will be delivered - the complexities of creating a debugger that will work across multiple machines are many.

Can you be more specific about what you want to do?

For example, a server job with calls to routines to run other jobs is a possibility. But it may not offer the snappy performance that web clients typically demand.

Courtesy tip: do not hijack threads with unrelated questions. Start a new thread.

I did not mention the Pivot stage. The Pivot stage does not support vertical pivot. The PXPivot stage will, in a near future release, but does not currently do so. I provided a solution that does not involve the Pivot stage, but shows how to accomplish vertical pivot in a parallel job, which is what...

Yes, vertical pivot as you identified in your subject line. First, partition and sort your data by the key column. Use a stable sort, so that the original order of rows is preserved for each key value. Use a Transformer stage to build lines of output containing each new value in a new field. Use sta...

So that's not the same problem, and therefore warrants a new thread.