DSXchange

Performance will probably be worse, because hashed file I/O from routines can not use the read cache. Lookups will be performed at disk speed rather than at memory speed. That said, to answer the question you asked, one routine is better than many in this case. The routine code itself will be reside...

You support provider can supply the Orchestrate manuals at no charge for the asking.

You are doing three separate lookups to Hashed File stages using three separate reference links. Each of these has a NOTFOUND link variable associated with it so, in your Transformer stage, you can direct stream input rows to a separate output based on the constraint RefLink1.NOTFOUND Or RefLink2.NO...

"Out of memory error" suggests that you don't have enough memory.

Please post the generated OSH so we can see what you actually designed. The message suggests that at least one mapping is missing from the Transformer stage.

Do you have Runtime Column Propagation enabled?

DS.CHECKER is the Cleanup command. It has not found any undeleted, redundant or orphaned job or shared container files. You can force-delete your job if it does not contain any shared containers. Can you please identify the records in DS_JOBOBJECTS using the following query? Post the results. SELECT...

Good catch. However it IS in the Orchestrate Operators guide. Page 13-19 in the version 7.5 manual. Curiously, string_trim does not appear in the version 7.0 Orchestrate Operators guide.

Welcome aboard. :D

Did the PACK actually get installed in the other project? Don't assert, check.

If you want to use separation 1 (512 bytes/group) - which your calculations did not use - then simply change the bytes/group in the formula. HASH.HELP is a very old utility, and does not factor in 64-bit pointers. The new calculation would yield 5,913,830 groups again factoring in 20% overhead (whic...

Let's review. It used to work. Now it doesn't. Nothing's changed in DataStage. Some DB2 patches have been applied. WHY do you think there's anything you can do in DataStage to remedy your situation? Go back to the vendor and complain that the DB2 patches have broken DataStage, and get them to supply...

Welcome aboard. :D

There's nothing out of the box to perform this particular task in parallel jobs. It's probably easiest to use something like an External Filter stage invoking tr sed or awk to perform the parsing.

Refer to my answer to that post.

Your time would be much better spent doing other things. Benefit from dynamic hashed file technology!

No that is not what I meant by a "here script". Read my posts carefully again, and get some advice from someone competent with shell scripts.

You're going wrong by not reading carefully what HASH.HELP is telling you. It has begun by examining average record size and coming up with a separation of 1 (that is, a group size of 512 bytes). It's recommendation for modulo is based on that group size. Your calculation was based on a group size f...

Unless the data are absoultely fixed width in every row it's impossible to calculate accurately. Use a storage overhead of 14 bytes per record plus one byte per field (for the delimiters) and round up to the next multiple of 4 or 8 bytes. Remember that there are no data types; you can't assume 4 byt...