DSXchange

Thanks, I'll have to check them out. Interesting, but my understanding is that the XML Output stage is based on Xerces - perhaps the others as well.

Let's see... 9,000 source records... 2 seconds per query... 18,000 seconds... 300 minutes... 5 hours. Why would you think a 7-8 table join lookup, accessed directly from a database with the ODBC stage, fired once for every row coming thru the job would ever be performant? Why not build a hashed file...

Hey Andrew, care to pass along the name / site for that external XSLT parser you mentioned?

kcbland wrote:Methought there was a repair repository files under the DS.TOOLS repair indices processing? Was I drinking heavily at the time?

Possibly, but that's a story for another day.

Convert the first day of the parameter year to internal format and then increment by 1 until the year changes.

Install Access.

Agreed, but the FMT solution seems... better. At least from an educational standpoint.

Try just fmt(right(In.Column,10),"15'0'R") as it looks like Arnd misunderstood the 'tacking on five zeroes' part.

Well, one of the two of us did.

Two schemas in the same instance? Sure. Two instances would require dblinks.

Many answers. A search on 'sdksequences' will turn up much discussion on the subject. As to a code sample:


Obviously, you'll need to make a couple of substitutions.

ray.wurlod wrote:From memory it's cataloged as "DSX.UTILITYRUNJOB"

Sounds right to me.

I believe that Ray would like you to post the actual results (copy/paste) of each command rather than paraphrasing them.

I'm sure the documentation would discuss how to set this stage up for each db vendor. Have you connected to Sybase via DataStage before this? Do you have the Sybase OC installed on your DataStage server?

You're better off upgrading and using the Stored Procedure stage.

You can't use an anonymous block there, so the syntax is either 'exec' or 'call'. You will find that not all procs take kindly to being called like that and the only recourse is the use of the SP stage in those cases.

By specifying Checkpointing in job sequence properties, if we restart the job. Then job will start by skipping upto the failed record.this option is available in 7.5 edition... Um... no. If you restart the sequence it will 'skip up to' the failed job and restart the sequence from failed job. i want...