Originally posted by: TarunB
Garrett
Here are some ideas off the top of my head:-
1)
You could generate a DB trace for each call, and reading this file after every execution. The trace file will contain the word UPDATE or INSERT depending on the operation performed. Of course, you'll need to make the trace filename unique to avoid single-threading the map. And you'll need to have a separate map to process each trace file.
2)
Perform this update/insert via a stored procedure and have the procedure return a value depending on the action performed by the call. You can then analyse the value directly after the call from your map.
3)
My personal preference would be to do the 2 calls. The first would be a simple Select statement, something like "Select 1 from table where blah blah...", to check for existence of the row. And the second would be your update/insert. If you set the config up correctly, the connection will be shared and so there is no perceivable degradation of performance.
Let us know how you get on,
Tarun
#DataExchange#IBM-Websphere-Transformation-Extender#IBMSterlingTransformationExtender