DSXchange

You didn't execute the dsenv script.

The inserted tsort operator does not do any partitioning.

(Auto) does not guarantee hash partitioning. In fact, it's possible (probable) that it will choose a keyless algorithm, such as Round Robin, or Same. It may be possible to tell from the score which algorithm was chosen.

1 = run OK
2 = run with warnings
3 = aborted

Change your logic for failure from $rc -gt 1 to $rc -gt 2

I knew that; I was trying to make a point. Not everyone uses these utilities.

Not good enough!! Burn the midnight oil! This is urgent - it's required ASAP!!!

I missed the fact that you're doing a lookup.

It's a DataStage command, so it's in one of the DataStage bin directories.

$APT_ORCHHOME/bin/dsrecords

You can also use the Data Set Management utility in the Manager client.

It does. If you don't, DataStage will insert tsort operators to ensure that they are. You can see this happening if you dump the score. Identical key-basd partitioning guarantees that every key can find its buddy. (You can also use Entire partitioning on the reference Data Set to achieve the same go...

Who is "u" and what is "r"? Is "wantt" related to "Gantt"?

Please use English on this forum.

All you need is a copy of SQL Server for UNIX, and a copy of the Plug-In Writer's Guide manual, and you can fulfil your desire!

The second of these can be had for the asking. The first, however, does not exist.

DS_JOBOBJECTS contains only the design-time default values of parameters, not the run-time defaults (set in Director) or the actual values used in a run.

Did you Search the forum for SIGKILL? Others have experienced this in the past.

Primarily by avoiding the overheads in not having to transmit those 52 bytes for each of the millions of rows selected in a query. Especially if a network is involved, and they have to be put into and removed from network packets. Secondarily the saving in disk space; roughly 52MB for 1 million rows.

There's also a readme file that's made accessible on the final screen of the install process for the server. It also has "what's new in this release" section.

You need to change the trigger on the Routine activity so that it passes good return values and fails on bad return values. You will need to define your own convention for "bad" - perhaps null or less than zero.