DSXchange

Two things to enable one feature: 1. An After-Job-Subroutine equivalent for Job Sequences. Key feature - it runs no-matter-what when a Job Sequence finishes. 2. DSGetJobInfo(SeqHandle, DSJ.JOBINTERIMSTATUS) must return meaningful values for Job Sequences, rather than 99 (No Running). These two thing...

This certainly add weight to Ken's assertion doesn't it? I'm starting to think someone's been slipping LSD in my Corn Flakes. I really, really, wish I had a 9i database now so I could put myself out of my misery. There are still two threads of hope for me: 1.It seems a very odd error to get for non-...

Hey, you never know! There is a forum for enhancement requests, so I was wondering if maybe this landed here by accident. And since you really didn't ask anything like a question that I could see... It was cleverlly hidden amongst the verbal diarrhoea :wink: Is there any similarly elegant facility ...

Craig, be assured that I hold this forum in the highest esteem, but my expectations don't include alterations to the DS kernal. I'm having trouble accepting that no-one else has tried this before. Or is this the reason ETLStats exists? If there is no elegant work-around, so be it. The help for DSJob...

DSJobReport() does a great job for a Server Job. You just whack it in the After-Job-Subroutine and a log is generated to disk with minimum effort. Is there any similarly elegant facility available for Job Sequences (which do not have After-Job-Subroutines)? I tried calling DSJobReport() from a routi...

I've never tried it, but could it be done with a Stored Procedure stage?
Pass the SQL Statement to the SP as a parameter, use dynamic SQL to execute it in a stored procedure, returning a cursor of known column names/types.

If you try it and it works, post a reply.

Create a Transformer with two Stage Vars: - Name = OddRow, Derivation = EvenRow - Name = EvenRow, Derivation = InLink.ColName In this way, EvenRow always contains the contents of the current row, and OddRow contains the contents of the previous row. Now, if you only output every second row, you'll g...

In Designer, hit the Help button, go to the Search Tab, and type in DsJobReport. The page returned describes how to use the routine.

It's the weekend, so I'm probably on delay. I know I'll regret this in the morning, but isn't ,,,"Pint, US liquid",PT,,,, a legal CSV record? The SEQ file stage would handle this easily delimiter set to (,) and Quote Character set to ("). I thought the optional quotes were there to ha...

Just an additional note. I found the same thing when I started, because we do everthing from inside DataStage - we don't set up any aspect of the environment externally before launching DS. $ENV and $PROJDEF both got their values from Administrator. If you set your environment variables from within ...

Now I'm really confused. I thought the problem was with OCI Load Automatic mode. Is it also happening with sqlldr DIRECT=TRUE? Can you confirm the following: 1. Oracle OCI Stage (non-Load) - OK 2. Oracle OCI Load (Automatic) - Error 3. sqlldr Conventional Mode - OK 4. sqlldr DIRECT=TRUE - Error Is t...

The accepted DataStage method of calling a stored procedure in the before/after SQL of an OCI stage is to use the CALL keyword. eg: CALL proc_name() We use CALL because wrapping the proc name in BEGIN/END does not work (at least I've never been able to get it to work). The reason it works is because...

First, I hope that if those sub-partitons are hash sub-partitions then there are fewer than 12 of them. Otherwise why the heck would you not just partition monthly? It's a far more productive use of the technology. I'll assume that the partitioned table is a target of the ETL rather than a source, b...

Oracle 9 or 10? From your post you're using 10, but I'm sure 8 and 9 have this restriction My test case was on 10.1, that's all I have at the moment. My last Oracle project was 8.1.7, where Direct Path Load over Oracle Net also worked. Prior to that I used it on 7.3. In my experiece DBAs jealously ...

The load files have to physically be on the same server as the database. Also, you must run sqlldr on that server, not invoked remotely. Not true. SQLLDR may run in Direct Path mode across Oracle Net using a file resident on the same (remote) server as where sqlldr is initiated. It is slower across...