DSXchange

What do YOU mean by parameterizing? How do you intend it to change for each run ?

Your question itself has the answer.

Partitioning affects the flow of records.

Think about it.

You can locate the nth record and read them alone using unix commands.

It is regenerate the kernel. Different from apt config.

Assuming your job is now
DataSet -> Lookup -> Copy

Can you replace the lookup with copy and see what happens.

Also I assume your CPU and disk are not 100% occupied.

Or join with a row generator and do a split the fields.

Nagarjuna is correct. Try it and you will understand.

Your requirement appears incomplete.

Who is be using your output? With so many pipes, the receiving party will have an heck of translation to do.

How many partitions? and how big does your partition get before locking?

Copy the job and replace the target dataset with a copy and see whether it finishes.

If not, we will have next round.

What does your log say?

In that case, you can append a job to the export and do a "import selected".

I am not asking you to bypass DataStage. All I am suggesting is to find which part of the activity is causing trouble.

Checksum is a name and not a fixed logic.

You need to be asking this to your business or source team of file.

Read only ?

What if you copy the production hashed file into development area and re-run the job post the hashed file creation.