DSXchange

You can create another project to re-align the jobs based on your needs.

This is not recommended as you will soon loose track of jobs when future changes come along.

Did you look into Pivot option ?

Will the number of columns be same in all input data ?

Copy the file (if arrived) into a common location. Thus if today's file does not arrive, you will have the previous one there.

Crude but simple method.

How did you deduce the error was from idoc ?

What happens if you remove idoc with a sequential file ?

Why not run the data through an aggregator and include 2 extra dummy columns.
1.) A newline character after the key field
2.) A '(count)' string after the record count field.

That must do the trick.

Send the message to cout

What have you tried so far and what were your problems ?

Is the provided example the full record or only the occurs clause section ?

Your requirement is bit unclear for me. Maybe that is because you are discussing more than one issue. How are you calling your Unix script from DataStage ? Is it a before/after routine or from sequence ? If it is from sequences, you can use convert function to change '\' to '/'. How are you using th...

Search for userstatus.

Why a routine and not a transform ?

You can make it into routine as

Blink stands for backward link.

Did you have loads of messages or did you run out of space ? A search for blink may help you.

That will and does work. It is a tried and tested method.

ArndW,

Are you saying that the solution works or not ?

You can try

You can use fprintf with stdout or stderr to achieve this.

There are loads of options
1.) Funnel the rows
2.) Before/after calls from job properties
3.) Sequencer calling header and trailer using execute command stage

Ultimately, all will be prepared separately and then combined together.