DSXchange

Why not sort the child records by parent, then process the child records and reference the parent record for each child, use stage variables to accumulate intermediate results, and send the final result out to an aggregator stage using the LAST derivation for all columns except the key columns? You ...

Maybe not a "relic" script, just a term for the site's demise and everything on it. Maybe a better term would be "lost treasures of sunken Atlantis". But then we can't call you Doctor Jones.

Do the jobs in different projects reference the same hashed file, or a copy local to each project? If separate copies, consider the size of the hashed file and its current tuning state. Consider clearing or removing and recreating the files in both projects to have a fair comparison.

Why Lara Croft Tomb Raider? Why not Indiana Jones? I seem to recall him pulling some relics out of some dusty catacombs.

I'd bet a box of donuts you're having a permissions issue. If users of different groups created jobs, the data is stored in one common internal repository file but the runtime component files are under the userid of the creator. You won't be able to compile or view logs etc of someone elses jobs out...

Welcome aboard, hopefully we'll be helpful.

The Help button just says this: User-defined SQL query file. Specifies that the data is extracted using the SQL query in the pathname of the designated file that exists on the server. Enter the pathname for this file instead of the text for the query. With this choice, you can edit the SQL statement...

Put the fully qualified path to the file in the USER-DEFINED SQL text box.

There is a Server side command line export, but I don't believe it's available back on DS 5.x.

I gave a confusing answer. If the jobs are identical (import from one, export to the other) and produce different results, then look to the Oracle Client. You've pretty much eliminated DataStage as the variable. Data is data, that didn't change just by being on a different DataStage server. DataStag...

My assumption is that this is the Oracle 9 plugin? If the jobs are the same source code accessing the same database but coming from different DataStage servers, the only variable left is the Oracle Client installed. This message indicates that the metadata in the job is sized differently from the da...

Didn't you take my recommendation to setup a new account with custom dictionaries and F-pointers seriously? The VOC is bloated probably due to years of accumulated junk, as well as dictionaries for each file is probably overloaded with unnecessary copies of dictionary items for formatting reasons, a...

If you have any simultaneous reading and writing to the same file you will have issues. Because your job has a syncronization point ahead of the output to the file, this doesn't necessarily occur. It's best to avoid your design anyway because you lose re-start capability, as you've destroyed the ori...

You probably have either an embedded carriage return/linefeed or a truncated row of data. Set the format tab appropriately and incomplete column metadata actions.

What are the source and target databases (DB2?). If DB2, try changing all columns with metadata DATE to TIMESTAMP and see if that fixes your problem.