DSXchange

I fear - without being able to check - that that's the way that mainframe jobs work. Try re-designing the job to check for the existence of the primary key, and directing your links accordingly. Only attempt an insert if the primary key does not exist; otherwise either perform an update or discard t...

It's not a "problem with performance" - it's a "problem with design".

Follow the suggestion that Klaus made and be pleasantly surprised - you are not unnecessarily processing (millions - 100) rows.

The secret of success in any performance tuning in computers is to avoid doing unnecessary work.

Etiquette Note We don't do "urgent" here. This is an all-volunteer site; no-one gets paid for posting. We post as and when we can. Some of us are asleep when your need is urgent, if indeed it really is. If you want urgent response, sign up with your support provider and learn the true cost of "urge...

Does Pervasive have the ability to dump its tables in text-file formats? If so you could script that, run the script from DataStage, then process the text file.

More likely would be a need to re-index DS_JOBOBJECTS. Best is to re-index ALL. You need exclusive access to the project to run DS.REINDEX successfully.

We tested the sql code on oracle DB its running fine. It seems their is some problem with datastage server. Any tips of how to find the problem? What has changed in the meantime? If the DataStage job has not been changed clearly it can not be the culprit. What is the total other load on the machine...

Could be a BASIC Transformer stage in a parallel job; however the OP specified server job.

Re-compile your job. There WILL then be a file in RT_BP124 that you can inspect.

Even though you know what you're doing, DataStage doesn't know that you know what you are doing, and still feels the need to warn you about possible conversion problems. You can not disable generation of these messages. You can, however, create message handlers to demote these messages to informatio...

The initial value must be obtained from somewhere, using select max(keyvalue) + 1 from table; which is a query that executes in sequential mode - no need for parallelism since it can only return one row. Do this in one job - perhaps even a server job - and capture the result into either a file or th...

Technically you did, through inaction, define the final delimiter and other properties. These have to appear in the schema file so that the record layout is properly understood. The mere fact that you chose to accept default values for these properties doesn't affect this need. However, record delim...

Lots of folks just lurk and learn - maybe too shy or not confident enough to post. I don't think we can assume that 0 posts = inactive. Perhaps >0 posts but none in the past three years would be a better indicator. What would be the most appropriate process - a private message and/or email suggestin...

I don't think you'd get one on Wimpy budget.

You can not do this easily in one job. A better approach is to run an earlier job to get the next key value, then feed that as a job parameter to your main job. Both jobs can be run from a job sequence.

If it's never been run before, "append" will create the new file. Otherwise you can include a Nested Condition activity to determine whether to truncate the file before running the remainder of the Job activities. This will have two outputs (yes and no); the yes goes into, say, an Execute Command ac...

Why?!! You want to rewrite what DataStage already does? Why did you buy DataStage, then? There is a function library called the BASIC SQL Client Interface where DS BASIC routines can act as client programs to database servers, but it has one or two disadvantages, particularly the licensing cost of a...