DSXchange

The object manager referred to here is an OLE Automation server, perhaps dsobjects.dll. If this has become corrupted in memory you must invoke the "Microsoft solution" - re-boot the PC. If it has become corrupted on disk you will need to uninstall the DataStage client, re-boot the PC, and then reins...

Make sure that the deadlock daemon (uvdlockd) is running.

Make sure that the job finishes reliably - even an after-job subroutine that performs a small innocuous task.

No. Read my third paragraph again.

Of course. But first describe your rules for obtaining the median from a partitioned (parallel) data set. Do you calculate the median from each partition separately, then form the average of those medians. I can save you some time; that approach is statistically invalid. Nor is the median of medians...

... except that day number is zero-based (what isn't?)

So you will need a slight adjustment on Craig's Field() function. You can decide which day is the first in the week, which will cause the list of day name abbreviations to be different, appropriate to your locale.

Change your mindset. Parallel job philosophy is "one stage, one task". The Lookup stage performs lookups. It has no other function. If you want to pre-set the lookup key, do so in an upstream Column Generator stage.

The file that you specify in the Data Set Management utility is the one whose name ends in ".ds". Not the long names of the data files themselves.

If you are using this sequence number to insert new rows into a table, why not retrieve it from there? Have a database stage on the reference input link to the Lookup stage. SELECT MAX(keyvalue) + 1 FROM targettable; Make sure you partition the reference input to the Lookup stage as Entire, ...

You do NOT require a string, because the weekday name is not part of a timestamp.

You need to extract the date part from the timestamp, then the weekday number from the date, then the weekday name from the weekday number, as I described the last time this question was posted.

The link row count shows the number of rows sent, not necessarily the number of records successfully loaded.

That's why we have reject handling.

OK, I had to check those things first. In parallel jobs, where data types are strictly enforced, the data browser shows complete precision and scale for decimal values, adding leading and trailing zeroes as required. I determine from that that the field in question in your example is defined as Deci...

What's the problem with Usage Analysis? That's what it's there to do.

Have you not taken appropriate pains with your metadata management?

Even you may be asking about installing Version Control. It is separately installable from the DataStage CD.

I dunno. The "corn" scheduler would explain many things...

Is there some reason you are not using the QualityStage plug-in stage?

What is the second -ipe.env option intended to achieve?