Page 1 of 1

Volatile table usage in Teraradata Connector Stage-DS 11.3

Posted: Wed Apr 22, 2015 4:11 am
by bgs_vb
Hi ,

We are going through jobs migration from Datastage 8.1 to Datastage 11.3 . But we have observed one issue in Teradata Connecitor stage behavior between these 2 versions :

As of current design ,we are using Teradata Connector stage with setting as
Access method-Bulk, Transaction Mode-ANSI,Character Set -ASCII .

We are using before sql statement and creating few volatile tables (ex craete volatile table p1_table ............) which are then referred in after sql statement .

Job was running fine with above setting in datastage v8.1 but in Datastage 11.3 ,it is failing in after sql statement stating p1_table(created in before sql statement) doesn't exist .

I am suspecting ,its something to do with sessions ,but not sure how to fix .

Request your help in resolving this issue .

Posted: Fri Apr 24, 2015 10:21 am
by priyadarshikunal
What is the error? Are you using same user to connect to database and is it the same database you are connecting from both environments?

Posted: Fri Apr 24, 2015 10:30 am
by chulett
Sounds like a support question to me.

Posted: Sun Apr 26, 2015 11:39 pm
by bgs_vb
Yes we are using same user to connect to same database in Before SQL and after sql .I also tried to retrieve session if in both "Before SQl" and "After Sql" statement ,and it happened to be different session id i.e Terdata session is disconnected after execution of before sql and then connected again while execution of after sql .So It is not identifying volatile tables created in Before sql Statement phase when referred in After sql .

However in Datastage 8. x ,also I tried to retrive session id in both Before SQL and After SQL ,and it happens to be same .So the same job is working fine in Datastage V8 .

But hwo do we fix this in Datastage 11.3 ,are there any environment level parameters or any patches or any other settings avaialble ,which could be set to fix this issue ? There are 400+ such jobs and changing datastage job will be very costly .