DSXchange

HP/UX perhaps? If so, switch from ksh back to sh and the core dumps should stop. There were also some old posts about setting LD_PRELOAD, I believe... but first let us know what O/S.

You can also schedule a little piece of job control code rather than the job itself and have the job control do all the normal bits - attach, prepare, run, wait, check status. Afterwards, do whatever needs doing if the job aborts.

Read each record in as a single large varchar. Use the 'count' function to see how many '~' characters are in each record. Use either @INROWNUM or the 'field' command to peel off the first value in each row to know which record type you are processing and how many tildes it should have.

ArndW wrote:Although some people (who I won't name although their handle is chulett) think that Windows are a "toy"...

Missed this last week. Actually, I think you are referring to Uncle Ken here, Arnd.

falchetto wrote:When executing a batch in datastage

You might also want to explain what this means. A Sequencer job that runs other jobs? A script from inside a DataStage job? Old style hand code in a 'Batch::' job? What?

What about 'sendmail' from the UNIX server? Oh, and let us know what O/S flavor and version you have there.

My first assumption is that the server is running out of resources and is unable to spawn new processes associated with the running of the DataStage jobs. Yes, swamping your server is the typical cause of this error. How many of these 'child sequences' are you running at once and how many jobs do t...

Jeez... you guys should get a room.

Not strange at all - you should make no assumptions about the environment your commands run under when launched from a DataStage job. The fact that the command works under the same user from the command line is fairly meaningless. Other than to know it actually works that is. If you had posted the r...

Post the syntax you are using for the command and the complete error from the job's log.

Documented in the BASIC manual, and you have to edit the DSParams file so that it knows they are legit functions. A quick search turned up this post from the Rayman himself with the details.

Try switching to the 'String math' functions and see if that makes any difference - SADD, SDIV, etc off the top of my head. They will show red in the derivation when you use them unless you take some extra steps, but don't let that bother you right now. They are there to handle large numbers, see if...

Welcome aboard. :D Sorry, many questions... Do you have any idea what this 'special character' is? Can you post some examples of input and output values? While the datatype in Oracle is a NUMBER, how are you defining it in DataStage - type, precision, etc? Do these numbers have more than 15 signific...

Hopefully that fixes it, as they should have all shown up - and a blown index would go a long way towards explaining the behaviour. A tip for creating Sequencers that I had no idea you could do until I saw someone else do it. Instead of adding an 'empty' Job Activity stage to the canvas and then sco...

With emphasis on the word extreme by the way.