DSXchange

Take a backup first. Export your projects. Do it twice!

Axiom: things only go wrong when you don't have a backup.

Talk with your DataStage owners - you need access to the DataStage manuals. Even I need access to the DataStage manuals occasionally (tell them that!). It would not be difficult for them to put them on a shared drive.

There is no IN operator in a Transformer stage. You could - in this particular case - use an Index() function. which returns 0 if the value is not in the string of digits.

It's a BASIC routine that could be invoked from a Routine activity's routine, when pre-processing the file. And, yes, the length is stored at the beginning of a string. Len() is an extremely efficient function therefore.

No. There is no such thing as a "hash" stage in either edition. Server edition has a Hashed File stage. It is a passive stage. Open its properties in the Stage Types branch of the Repository to verify this assertion. Parallel Edition has no Hashed File stage. The closest thing is a Lookup File Set s...

You may have an indexing problem in the DS_JOBOBJECTS file, or a corrupted job design.

Operating system commands can not correctly delete hashed files that were created in an account (project), without returning to the DataStage shell (which seems wasteful).

No idea. What are hash_1 and hash_2?

What error do you get when you try to run the job? Acually, this probably needs to be a new thread.

FUNCTION Trimmer(TheString,Delimiter) If Delimiter = "" Then Ans = TheString End Else Ans = "" * Remove surrounding delimiters, if any. If Left(TheString,1) = Delimiter And Right(TheString,1) = Delimiter Then Temp = TheString[2,Len(TheString)-2] E...

Good catch. But a routine that processed a character at a time with lookahead would do it easily.

If they have, they've chosen not to post it here. Obviously using a different quote character is feasible, but the script that changed them would need to be careful (it could actually be easily done in a routine, using Trim(TheString, '"') , but a filter command works differently). Hey,...

Who imposed that restriction? The only way you can do it in a "single server job" would be for that job to contain/invoke a routine. How you delete the hashed files (note: not "hash" files) depends on whether they were created in an account (project) or in a directory.

Define "performance" in an ETL context. Only then might one be able to address whether whatever it is can be made "better" (whatever that is).

Please do not cite rows/sec as your performance metric. Search DSXchange for many rants from me about why it is meaningless.

All is well. The date format you get when you view a date in the View Data browser is governed by the default date format string currently in force. The date itself, once it's been converted, is a binary form.