How to capture number of rows processed

midmaxds · Post by **midmaxds** » Mon Apr 25, 2016 6:35 am

Is there a way we can capture the rows written into target (file/DB) &
write that information to an audit table.

For example if we have a job like this

Sequential File --> Transformer --> DB2

Say 1000 records are written to oracle DB. I want to insert the value 1000 into an audit table (which is in DB2 but another schema).
We are not planning to use parallel/Server routines. Any other alternatives please?
We are running on 2 nodes and DataStage is in UNIX.

chulett · Post by **chulett** » Mon Apr 25, 2016 7:50 am

So you want to capture the number of rows processed without any evil routines? Add an aggregator from the transformer, count the records and write that to your second target.

IBM Analytics Champion 2009 - 2020 · Post by **asorrell** » Mon Apr 25, 2016 4:52 pm

Craig's methodology is good as long as you just need approximate counts. However, any rejected records would throw the count off.

If you have to get an exact count that excludes rejected records, you might want to check out the DataStage Operational Database, to see if it is stored there. If not, then you'd have to do the routine calls you are trying to avoid!

chulett · Post by **chulett** » Mon Apr 25, 2016 7:20 pm

Rejected records? We don't need no stinking rejected records!

midmaxds · Post by **midmaxds** » Wed Apr 27, 2016 5:46 am

Hi Craig and Andy,
Yes, the number of rejected records are not needed at this point.
So, we will go with the aggregation. Thanks much.