You are right. I was bit confused when i read his very first post. But subsequent posts i understood that he is considering insert and update both as a single transaction. So i suggessted him to use 'enable transaction grouping'.
I do not think it is possible to rollback the changes when you use 2 ODBC stages. It would be treated as two different transactions and there would not be any link between these two stages.
I still dont understand why even with singl link to table and transaction size set to 0 if i have warning on first row and stop the job from job sequence all rows after first are inserted/updated on database
And how to do rollback with singl link ? Doesnt matter two or more
The job has to abort in order for any uncommitted rows to be rolled back. And with a Transaction Size of zero, then 'any uncommitted rows' = 'all rows'.
-craig
"You can never have too many knives" -- Logan Nine Fingers
Contemplate adding a "run number" or timestamp to the target table. This will make unwinding any committed rows straightforward, albeit through another transaction for the DELETE statement.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.