commit point when loading SQL sever DB using ODBC enterprise

bmadhav · Post by **bmadhav** » Mon Aug 22, 2005 11:42 am

I am trying to do inserts into SQL server database using ODBC enterprise stage fron a PX job. I have to insert approx. 30 million rows of data into the SQL sever database.
My input is a sequential file on AIX. The database is SQL server 2000.
How do i set the number of rows to commit/transaction size?
When i am running my job, it looks like it is commiting after row being inserted.
I did look for APT* env variables. I do not see anything specific for the ODBC enterprise/SQL server stage.
Anybody run in this situation before?

Thanks
BM

ray.wurlod · Post by **ray.wurlod** » Mon Aug 22, 2005 4:39 pm

Welcome aboard! :D
From memory (I'm not presently working in the EE environment) there are two generic environment variables you need. One is APT_COMMIT_ROWS (or something like that), the other is APT_COMMIT_INTERVAL.
Check the Parallel Job Developer's Guide for full details.

bmadhav · Post by **bmadhav** » Mon Aug 22, 2005 6:47 pm

Ray

I did check all available APT variables at the project level and did not see any generic APT_COMMIT*.
I will dig into the Parallel dev guide and see if i can find anything.

Thanks
BM

ray.wurlod · Post by **ray.wurlod** » Mon Aug 22, 2005 11:48 pm

They're not all in the Parallel Job Developer's Guide - open the Administrator client, select your project, click Properties, then click Environment...

You should find them in that list of environment variables. I was using 7.5.1A last week, and they're definitely there.