DSXchange

On an upsert it's highly unlikely since only individual record-level locks are taken, and these should be independent in the different streams. I was making the point that stored procedures can contend for the same resource; if the same procedure is invoked from parallel streams (multiple processing...

When you format it to two decimal places it regains its leading zero.

These are configurable using optionally added properties. The default values may be found in the Parallel Job Developer's Guide manual. Are you using a Difference or a Compare stage?

Welcome aboard. Hash files are used for shaping blocks of hash. The term you need is "hashed" files. These use a hashing algorithm to determine the physical location of records, rather than needing a table scan or index scan. This is the fastest key-based access method there is. When the structure i...

When this return message is printed, the job interprets it as a failure condition and aborts. Is there anyway to ignore the print message and prevent the job from aborting? No. That's the documented behaviour for any programming using the ODBC API. If SQLNumResultCols() does not return a value, the...

The routine exists; it's in the SDK and it's called UtilityRunJob. If you don't like exactly how it works, clone and customise.

Unless, of course, you want separate messages from each of the jobs. IN that case, each job requires its own "lack of success" trigger to its own Notification activity (and maybe a Routine activity as well to invoke UtilityWarnToLog).

Tigger should only be used when you want something bounced.

That generates a random number - not, I suspect, what the OP desires! In server jobs you can simply perform arithmetic (for example add zero to the number) and it returns the shortest possible string. You could then format it to have two decimal places. In parallel jobs, it is not quite so easy. You...

ODBC Enterprise stage.

For those on Windows servers, there is also a SQL Server Enterprise stage.

... or it might be a "self deadlock" caused by trying to run the same stored procedure from multiple processing nodes, each trying to lock the same resources. Your DBA will be able to tell you if that's what it is. If it is, you will probably need to run that stage in sequential mode.

True parallel multiload from DataStage PX has been achieved by an Ascential pre-sales engineer in Melbourne Australia during a Proof Of Concept. Alas, I did not write down any details at the time - will have to get in touch again to see how he achieved it. As I recall it was several hours of try thi...

skrakesh wrote:The problem is the environment being used here doesnot support sequencer. I need to loop within DataStage job.

All DataStage environments except mainframe support job sequences.

I seem to be busier than the proverbial one-armed wallpaper-hanger with a bad itch.

FIGJAM

It's not as hard as it appears. Create a job control routine via the Add Job button. That is the essence of running a job, albeit primitive: attach, set parameters, request run, wait, check status, detach. Job sequences use a "busy wait", and lists for keeping track of which activity is running, che...

Using back to back active stages is the preferred method of working in parallel jobs. Don't forget that there is a virtual Data Set (with buffering) between each pair of stages in any case. You can see this (the control files whose names end in ".v") in the generated OSH.