DSD.StageRun

Mark j · Post by **Mark j** » Fri May 14, 2004 7:39 am

Hi
When i am running a job which has 4 lookups and a driver file with 30 million records and 5 output files and each of these files are directed to seperated directories then why while running the job, When i looked into project directory this file is increasing in Megabytes as processing for job is going on.... &PH&/DSD_StageRun1392 .... my worry is our project directory is only 7 GB and if i calculate for this single job its taking up upto 4GB when the job is finished. I have checked for all paths and i don't see any wrong with them(nothing is default) so what could be the reason for this and my job is processing at 42 rows/sec which i believe will take around 20 hrs to process allthe records.

Thanks
Mark

kcbland · Post by **kcbland** » Fri May 14, 2004 8:03 am

The &PH& directory captures DS BASIC runtime error messages. If this directory is filling up then it's a sign that the job is generating zillions of error messages into the screen capture file you noticed.

My suggestion is to run only a few rows thru and see the error message and correct it.

Mark j · Post by **Mark j** » Fri May 14, 2004 8:48 am

Hi Kenneth,

You are right , i am getting error messages due to handling of empty strings but why do these errors doesn't show up on the job log..

Thanks
Mark

kcbland · Post by **kcbland** » Fri May 14, 2004 8:52 am

Because screen output are runtime error messages and DB rejects are responses from targets.

ray.wurlod · Post by **ray.wurlod** » Fri May 14, 2004 4:22 pm

DSD.StageRun is the name of the subroutine executed to implement a Transformer stage.
You can examine the generated source code in the RT_BPnnn directory, where nnn is the job number from DS_JOBS (search the Forum for RT_BP).
You should be able to diagnose where the problem lies. The fix is going to be to correct the handling of null values and zero-length strings.
If all data handling is correct, none of these warnings will occur.