DSXchange

You do not have an ELSE clause if your OPENPATH statement fails. In that case, you never assigned RecordCount a value.

Okay, I thought you were using just sequential files. Well, write a DS function to COUNT your hash file and interrogate the results. Put that into the BEFORE-JOB activity.

To use UV/ODBC to aggregate a hash file, you need to setup a pointer so that DS recognizes your hash file as a Universe file/table. Search this forum for SET.FILE, you'll see it's been covered. It's unfortunate your data is in a hash file, otherwise, you might consider dumping the hash file to a seq...

Write a ksh script that returns a non-zero returncode if a passed filename has zero rows or does not exist. Put that ksh script into the BEFORE-JOB activity of every job that needs to fail on a zero-byte or non-existent file. For the parameter to the script, pass the fully-qualified path and filename.

Search the forum for using a macro. It allows you to have a dynamic filename without involving job control. However, I don't do what you're doing. I prefer to have jobs write to a constant name with a variable path. At most I'll parameterize a portion of the filename, like: #WorkFileDirectory#/D_CUS...

Or take option #2 and in the before routine run a script or function to count the file and if zero then blowup the job. I tend to like things a lot more graceful than that, so options #1 and #3 are pretty good. I'd prefer #1, because why run the job if you don't have to?

Well, you can't do what you're doing. You have to either have a job control process check the file and decide if you want to run the job, OR, have that job control process run the job and then analyze link statistics to determine if the job performed as you wished.

A job cannot change its own job parameter value. Job parameters are static values once a job has started.

nag0143 wrote:Result = DSGetLinkInfo(DSJ.ME,"stagename","linkname",DSJ.LINKROWCOUNT)

DSJ.ME references your current job. You should be using the job handle for the job you want to query.

Make sure your filename is like "/wherever/#RunQual#".

Just use the hash file stage. Why are you using UV/ODBC???

Try logging the Result value and see what it contains.

Under your Start menu, Ascential DataStage, Online Documentation, is a wealth of documentation in .pdf form. You will be stunned at the level of documentation there, on all of the client tools. In addition, you can get the Tutorial off the client install cds and learn from there. This forum is not a...

Different options:

1. Use job control and check the file first, rather than run the job.
2. Put in a BEFORE-JOB/TRANSFORMER check of the file first and have that routine return an error exit code or log a fatal message.
3. Use job control and the link statistics after the job has run.

A UV stage has two purposes: (1) native connectivity to IBM's Universe database, of which DataStage Server is build on a proprietary derivative; (2) SQL interface to a DataStage hash file.

You can read this for more info:
http://www.dsxchange.com/viewtopic.php?t=85364