DSXchange

Mea Culpa, I am using the DB2/UDB Enterprise stage in PX and the attributes there are different. I realized that one can still use the DB2/UDB API - and that is what you had specified. Sorry. If you don't specify the "create table action" then the code won't get executed; I wrote a test job and my e...

Nitin, OK, then I do have a question and ask that you check your job to confirm that you did in fact set "insert without clearing", the array size and transaction size. Please note that the DB/2 PX stage does not have the attributes, there is no "Write Method = Insert". There is an Options -> Row co...

If you used "insert without clearing" then you modified a server job, that doesn't exist in PX. Could it be that you are modifying one job and looking at the log of another (I've done that before... more than once)?

Ray - those attributes don't exist in DB/2 PX stages, and the giveaway was "insert without clearing" since that only exists in the server DB/2 stage. Just as you stated, the log entries don't make sense in this context.

This is a SERVER job! I just realized that we have a bit of a mixup here. Your initial post specified a number of stage attributes that don't exist in PX but I've been looking at this from PX angle - you did post in the PX forum! So, as a Server job: 1. What did you set as the "Create Table Action"...

Does anyone have any information on whether or not a direct hashed file stage is on the horizon for PX? I realize than one can go through ODBC to access UniVerse/DataStage hashed files, but would be happy to see a direct stage to read from and write to hashed files in PX jobs. I think that incorpora...

Yes, but with limitations. Read up on the BASIC transform stage in the documentation.

If you recompile and force compilation does that change things? It might be an old binary image that (for some reason) isn't being updated to reflect the new source code. An alternative is a save-as new job and compile/execute that.

In all the time that I've been working with DS I've never seen an error in this part of the code. The system variable is autoincremented for each row and is completely independant of the data being processed. I think both Ray and I are saying that it is very unlikely that the error is happening with...

Rafik,

the basic approach is correct, but something seems to not be working correctly in your case. Could you also output the change_code value and see what it is for each row?

I would look elsewhere before assuming that the @INROWNUM builtin transform value is not functioning correctly. Could you perhaps explain a bit how your job looks and how you determined that this error was happening?

That looks sufficient - have you tried it?

Instead of If JobName = DSGetJobInfo(DSJ.ME, DSJ.JOBNAME) Then End Else hJob = DSAttachJob((JobName), DSJ.ERRNONE) JName = DSGetJobInfo(hJob, DSJ.JOBNAME) End I would just use: If JobName = DSGetJobInfo(DSJ.ME, DSJ.JOBNAME) Then hJob = DSJ.ME ELSE hJob = DSAttachJob(JobName, DSJ.ERRNONE&...

You added the parameter, but what is the value that you used? The manual says the default is 2Mb but your error message states that you didn't have 4Kb --- that's a huge difference in values!

In the job log for the new job does it state that only 10 records were processed? Does the same thing happen if you enter "/dev/null" as the destination for the target sequential file? If you manually create a new test file with 20 records and no data and similar keys can you save & retrieve all...