DSXchange

Craig

Thanks. It worked. So does that mean it's not able to allocate the memoey specified in the buffer size ?

Thanks

I have a Job which is like this... Hashed_File Hashed_File | | | | SeqFile---------------------------------------SeqFile This job runs in 2 other environments and fails in one. Error message is... Abnormal termination of stage JOBNAME.TRN NAME detected JOBNAME.TRN NAME.LINKNAME: ds_ipcflush() - Erro...

Craig - Yes. I got that message also as Info next to the warning while reading from the hashed file...

Ray - So If I have this warning in my log file anytime then I can just ignore that (I understand that it slow down the performance)...Is that Correct ?

Thanks

OK..I know that this topic is covered here many times. I read some of them and am trying to double check that my understanding is correct. My job design is as follows...Both source and target is Seq Files. I have say 7 Hashed file lookups while a record goes from source to target. I have the 'pre-lo...

Than It contradicts what Ray is saying..is that correct ?

Thanks

I did this in one of my Hashed file...

od -x DATA.30 | head -1

and I got this...

0000000 acef 020c 0000 001e 0000 0000 0000 0001

Does this means this hashed file is 32 bit ? But when I look into my uvconfig file I see an entry like this...

64 BIT_FILES 1

what does this mean ?

Thanks

I have a job which has 10 look up's which was running slow so I enabled the option of Pre load file to memory in hashed files and it started to run fast...These files are already built when this job is running. My question is I have another job which has 17 lookup's If I do the samething on that job...

wow...that makes the job easier...Thanks Arnd...

Can anyone tell me an easy and efficient way to add a column to an existing hashed file ? This has file has 14 million records so it will take atleast 8 hours to reload the hashed file with a column. I am looking for any command that can just add a column to an existing hashed file and the value of ...

The way of echoing env variables to capture them in Routines is making the job run really slow which make sense as every row it's going to Unix and coming back. I think I am better of changing the Routine to add new parms so that I can pass env variables as new parms. Initially I did not want to do ...

Craig - I don't know if that' the case. I am able to see the results now even when I just test the routine. I don't understand what's happeneing. But now I see output for the same routine. The issue I was facing before also solved when I changed the code from this HashFilePath = Trim(UnixOutput) to ...

Sorry for the confusion. we share the ID. I replied when I try to view the response then realized that the responder may look for OP's name in the post so I pasted it again... Now...coming back to the question...Variables are defined in the DS Administrator. Values are going to job log. But still I ...

What I meant by not working is that it's not returning me anything.
When I test the routine and see the result it's empty.

Thanks

Craig - I understand that. Can you tell me where I am making mistake.
Why my simple routine is not working ?

Thanks

It won't sipt out the value as the variable is DataStage environment variable. So Unix will recognize only If I run from DataStage.
Is my understanding correct ?

Thanks