DSXchange

Add an extra non-key column on which to pivot. It contains a constant, being a comma-separated list of your (literal) column headings.

I am trying to show my job design to handle the change capture. How can I paste/attach my job design in the forum/message. This forum does not support attachments of any kind. The recommended approach is to show your job design inside a set of Code tags. Use Preview iteratively until the layout is ...

I must disagree with Sainath that you can use OS permissions. In particular, most objects exist as records in database (Repository) tables, rather than as operating system objects. So any approach that uses OS permissions would be an "all or nothing" approach. DataStage does not have a source code c...

The DataStage logic is exactly the same as the "logical logic". If a field is a candidate for recording change, then a new record must be created whenever there is a change in that field. Therefore, you must determine whether a change has occurred. This involves being able to see (pre-load) what's a...

You might also benefit from reviewing the Lookups section in the tutorial that ships with DataStage.

sed (stream editor) is probably the fastest mechanism for removing empty lines, then pipe the result through wc -l

You can perform arithmetic on it. For example add zero to it. DataStage returns the shortest possible string when arithmetic results are presented. Other approaches use the Trim() function, the Fmt() function, or the Iconv() function, not to mention reinventing the wheel by writing your own routine....

Everyone's answered the question for transform functions, but none for (before/after) subroutines. The answer is still to add the "DSU." prefix, but the call mechanism is different, so no DEFFUN is used. There are two possibilities. The first is simply to add the "DSU." prefix in the subroutine name...

... of course, Arnd meant

Another possibility (which is really the Field function) is:

Or you can use the Compare() function and enforce left-justified comparison of the original strings. Make sure that each has the same delimiter character.

When you posted it was 12:21am in my location - I do demand the right to sleep occasionally! As you will already have found out, PX is much more pedantic with handling of metadata. You're trying to achieve something that was easy in server jobs, but is much more difficult in parallel jobs, that bein...

If there is a limit it's well above 22. Please post the error message that is generated when the job sequence won't compile; this will shed light on the likely cause. It's usually that you've omitted to supply one tiny but vital property value.

Type 1, Type 2, Type 2.5, Type 3 ... ?

As Lance said, there's been a lot of coverage in these fora; Search is your friend. Prefer "match all terms" or "exact match" when searching.

The File Pattern property allows you to specify a wildcard; the stage will read all files matching that pattern. So you could specify PA*.csv in the File Name property (or a more appropriate pattern to your needs).

My guess: Probably an artifact of the Lookup stage being incomplete (no reference input link or output link yet).
Do you get the same symptom with SeqFile ----> Peek ?