DSXchange

This answer has been posted before. Basically, the Sort stage gives more flexibility over resource consumption than sorting specified on input links in parallel jobs.

Try recompiling jobs. Invalid information may have been posted to RT_CONFIGnnn files while DS_STAGETYPES had corrupted indexes.

is close enough to account for leap years in most cases.

Make sure the value of RL really is 'Y'. Chances are that it isn't.

To get to that, right-click on a row in the Columns grid and choose Edit Row from the menu. Note that you can move to the next or previous column without leaving the Edit Metadata dialog. Note also that you must Apply or abandon any change before moving on: the dialog reminds you to do so.

You are WAY under-configured. Mine runs on a 1.3GHz CPU, has 2GB RAM and 160GB of disk space. And even then I would not use this in a production environment.

Welcome aboard. :D The Iconv() function itself does not calculate the date difference - it converts the date into DataStage internal format, which is the number of days from a particular day zero (31 Dec 1967). It follows that, if you subtract internal format dates, you get the difference in days.

In general, not within a single DataStage job.

You might build such a strategy into a job sequence; capture the results of the first query somewhere and use those results to supply job parameter values to a job that processes the table/view.

When's the interview?

I've seen something like this in a sequence of Merge stages, where the fix was explicitly to propagate the column (for example Rate_WA=Rate_WA). Clearly this is not feasible in a Join stage, but you might investigate the mappings a little more thoroughly. It's normal that key columns from the interm...

While it is true that you can use explicit Remove Duplicates stages, you can also specify de-duplication as part of sorting the input links. Check the "Unique" check box. And this can also be done with a Sort stage.

Put a Peek stage on each link in the job to help you to diagnose exactly where any problem is occurring.

An IBM class UV910 "UniVerse Theory and Practice" (formerly called "UniVerse Internals") is running May 01-04 in Sydney, Australia. While this is not directly about DataStage, I know that the instructor will be able to relate the UniVerse materials to DataStage. If you're interested please contact A...

Welcome aboard. :D

Something has gotten out of synch. You may be mollified to hear that this is quite rare. Probably best to clear the log in Director (Clear Log is under the Job menu), re-compile the job in Designer, and let it roll.

Methought there was a repair repository files under the DS.TOOLS repair indices processing? Was I drinking heavily at the time? DS.REINDEX is separate from DS.TOOLS. Without access to DataStage at the moment I can't recall if there is an option under DS.TOOLS to repair indices. Certainly they remov...