DSXchange

roceller, a) in the first words of his post that are visible to all users, Ray told you what the problem was. Yet you still complained, even after thompsonp went to length about the problem as well. b) I'll assume that you have been posting under another ID, since you stated that It's shocking that ...

Kumar - GRANT is a DataStage command. The error seems to revolve around those quirky Windows domain ids and I'm not at such a system right now so can't help - the command should have worked.

You need to use the GRANT command to effect this. Try GRANT DBA TO velag001;

Creating hashed file inside the project directory makes migration more difficult, in my opinion. If you always use the same path to create your hashed files then migrating is just a matter of copy that directory path and contents - if the hashed files are created in the project directory then you ne...

The job design information is written into several DataStage hashed files in the repository and the information cannot be accessed from the OS directly; even by using the builtin SQL and other query language much of the structure remains hidden because it is not documented by IBM/Ascential. The .dsx...

Array ought to be initialized with the MAT {ArrayName} = value command to make sure they have values.

Does your OPENSEQ have an ON.ERROR clause?

The array is stored in local memory of a process, and thus there is no possible way for different processes invoked at the same time to interfere with each other, at least when it comes to local or even global variable storage. Your error might be coming from the common routine which fills this dime...

Dmitry, fortunately you normally don't need to know all the options in this stage. If you are working with hashed files whose final size you don't know or files that will grow over time, then the only choice you should make is the type 30 or Dynamic hashed file. You can specify a minimum modulo if t...

You can query UniVerse Databases with either the native query language INFORM or you can use standard SQL queries within the TCL environment. Note that DataStage is no longer UniVerse but they are clones and very similar. The IBM website offers all of the UniVerse documentation free of charge in PDF...

There are too many factors that affect performance speed that you haven't told us for anyone to make a good suggestion. We don't even know how many physical processors your maching has and if the job is waiting on I/O or on CPU. And your other post only has one processing node defined, so perhaps yo...

No, but I did respond earlier today to someone at the same site and was able to check the job directly

The good news is that the error doesn't have anything to do with your job, it is a DataStage engine issue. The initial messages point to a corrupt index on the DS_STAGETYPES hashed file. I just checked to make sure and the 3 indices on this table should be fixed with a DS.REINDEX. If you try the com...

The error message is almost as useful as the one that says "a red light in the dashboard of my car is on" If you replace your XML stage with a sequential file output, does it run to completion? If you do a reset of your job, does your log file have something more informative under "from previous run"?

All of the RT_* and DS_* and SC_* are system files that might make the system unusable if you delete them. Unfortunately many developers use local hashed files in their jobs and those get written to the project directory. These are most likely the files that are clogging up your project directory an...

256Mb of RAM is less than what even the normal DataStage client recommends, to say nothing of installing a PX engine! I am sure that your disk is going run really hot because of all the swapping and if it runs at all it will walk or perhaps crawl.