DSXchange

I just wanted to confirm the source of the error, as the message doesn't normally relate to an Oracle write. What happens if you change the Oracle stage to do an UPSERT, does the message also go away?

Just for testing sake, can you put a PEEK stage into your job output instead of the Oracle stage and see if the error message persists?

It is possible to do this via access to the hashed file metadata contents; although it does take quite a bit of work. I would recommend you start with doing a project export into a .dsx file or, if you are more comfortable with XML, exporting in that format. The structure of the export file is prett...

Created a counted loop in your sequence that goes from 1 to a couple of million and let that loop with a command stage of sleep {n-seconds} in the loop so that you don't use too much CPU just looping if there is nothing for the 2 jobs to do.

What about doing a save-as of your job and checking to see if you can export that. If yes, then there might be some corruption in your project.

More tuning work has gone into the sequential file stage than the CFF stage. I think the limiter is the amount of CPU used in the CFF stage, since it needs to parse out all the columns and then interpret them. If that is the case and you have a system with more than 1 CPU you can increase throughput...

The reason I asked about timing was to make you check whether the dataset write job is complete before the dataset read job starts. You haven't confirmed or refuted that possible cause yet.

Unfortunately sequential read speed does decrease significantly as the number of columns increases. If you declare each row as just 1 column in a test job and write out to a sequential file stage what speed does it perform at (this will help give a baseline as to expected best case throughput). If m...

The DataStage BASIC programmer's Guide has a description of the BITXOR (V1,V2 ) function which you can use to revert your modified string back to the original one, given you know the string with which you XOR'd in the first place.

There are no hard-and-fast rules regarding number of nodes in the configuration file and number of processes - there are just too many variables introduced by how jobs are designed and coded. If you run your job several times with 1,2,3... nodes and chart the overall runtime you will see that perfor...

Could you explain which buttons/switches you are selecting in your export window? What happens if you export the job, source & objects?

Does your job "CopyOfTCIJBBKORFO" contain anything? It doesn't look like you selected a full project export.

Rajesh - to reiterate, the default value for a parameter will be used unless the parent sets it. Somewhere you are setting this parameter to a value. If you set the default value in the Job Activity stage explicitly then that value will be used when the job is called. I don't know what else to ask -...

Yes, the same error happens regardless of the number of nodes I use (1,2,4 and 8 for both the lookup fileset and test job).

I am leaning towards thinking this is a HP-UX specific issue and am awaiting a response from support (I hope that they have a HP_UX test bed available somewhere).

No, the output will be identical providing you don't change the lookup data while the job is running, in which case the sparse will show the new value and a normal one, which uses a memory image of the data as it was when the job began, will show the old value.