DSXchange

From what I recall, you need to use /SaveAsBmp=filename when calling it. Or in 7.5 I think they added a /R parameter to add a bitmap image of the canvas to a job report.

It all should be in the readme, I would think.

Welcome.

Make sure you set the Display Size properly in the Columns metadata as that information is what drives the stage. It sounds like you may have left that blank.

But - as noted - you can launch the job with the -userstatus option and then it (if set) will be returned as the exit status when the job completes.

Seems to be exactly what David is after.

Isn't there a -userstatus option for the dsjob command?

One other thing to note is that the Oracle stages work differently for dates than the ODBC stage... ODBC wants to have to the external format for its dates as input, while the Oracle stages use the internal DataStage format (when inserting via an Oracle stage e.g.) That would be the DB2 stage you a...

You are thinking of the deadlock daemon.

When you get right down to it, it's not about 'new' versus 'old' but - based on the fields you marked as keys - where any rows in the target database actually updated? When zero rows are updated that triggers the call to the insert... which can still fail because the 'real' keys already exist on ano...

Ah... but it will log the fact that the after job routine has been executed and the syntax used to do so.

If in the Designer you don't have an option to create Parallel jobs, you don't have the Enterprise Edition. And you're not the first person to be thrown by the online help, which includes every stage for all products. You'll find a Conditional Lookup in there as well, too bad it's only for the 390 p...

u can find this in the output stage, under properties tab -->Opiton ---> Clean On Failure Huh? Maybe in a PX job, but not in Server anywhere that I recall. The best thing to try is what Arnd suggested - reset the job after it aborts via the Director and see if any messages come into the log labelle...

kumar_s wrote:Cant we set that n to no limt, as we have in director.

-warn 0

(That's a zero, btw)

Where most messages go - in the calling job's log.

Well... first thing I'd do is look for example routines that ship with the product that do something like what you want to do. There are at least three in the sdk category that read from and/or write to hashed files: KeyMgt\KeyMgtGetNextValue KeyMgt\KeyMgtGetNextValueConcurrent Utility\UtilityHashLo...

What are we discussing here? A hashed file or a sequential file? Sequential files 'append' new records. You can't 'update' a sequential file. Hashed files update records based on the keys defined in the metadata and the new version completely overwrites the old version. New records get hashed in (ak...

snassimr wrote:How to do overwrite on second position in row when I faund my key in first position ?

Write the entire record again for that key with the new non-key values. Boom - updated.

snassimr then wrote:Its talken about seq file

I have absolutely no idea what that is supposed to mean.