DSXchange

Are all the servers running, and not paused? Is this the login to the deployment domain or to DataStage? What does your support provider suggest? Since it was working before and isn't now, what has changed? "Nothing" is not the correct answer, though the thing that has changed is not necessarily any...

Welcome aboard. Is this question about DataStage TX, or DataStage other than TX? If the latter what you request is not possible. If DataStage TX (which is suggested by the forum in which you chose to post) I do not know whether what you ask is possible, though I doubt it. The reason is that DataStag...

You have posted in a forum called DSXchange Testimonials, which is not the appropriate forum for a question about parallel jobs. Please post in the correct forum, where your question is more likely to be answered. You might also like to search the forum before posting, as this question has been ask...

You have posted in a forum called DSXchange Testimonials, which is not the appropriate forum for a question about server jobs. Please post in the correct forum, where your question is more likely to be answered. You might also like to search the forum before posting, as this question may have been ...

You only missed a word and a half! And unimportant ones at that. On that basis I've removed the Premium flag.

These days what a Transformer stage is, to all intents and purposes, is a BuildOp generator. Look at the generated code to see what I mean - though it's rather better documented than a lot of user-written BuildOps!

What can be done? Get IBM to fix the stage.

Parallel jobs can, however, be compiled in trace mode, which is very similar to server jobs' active stage tracing. You can even limit the rows processed.

What do you get when you type this commands at the UNIX prompt?


It's impossible to diagnose further without knowing that.

A more general solution requires you to realize that these stem from the use of Lookup stages. A Lookup stage (unless performing a sparse lookup) must perforce create a lookup table in memory, and then build an index (hash table) on the fly, before any lookup operation can be performed. That is why ...

How are you determining the number of records in the Data Set?

For someone to "take a look" surely they will need access to your environment, maybe even the mainframe on which DB2 is running.

Isn't this what you pay your support provider to do?

Note that, using the suggested technique, the same sequence of batch numbers will be generated on each processing node in your parallel environment. Is this the desired outcome?

1. Constraint expression (InLink.AAAA < 0) on output port #0. 2. Use stage variable to generate sequence and another to convert same to string which you can concatenate to your source string ("22222"). Note that you will need to take steps that you are generating unique numbers across all processing...

I'd have to agree, noting that you may also need quote characters (regular SQL rules) if the data type is Char or VarChar.