It seems that the Oracle enterprise stage only allows for setting the row or time based commit frequency at a job level. I have a job where I would like to set the main data output, which uses the write method, to as long a time and as high a row count as possible (using the 2 oracle environment variables whose exact names I can't recall right now) but I have another oracle output with far less rows where I need to commit after every row (this is used to update the load status and is used for restarts, thus needs to commit in order not to lose the update information if the job crashes).
I couldn't think of a way to get this done using just the Oracle Enterprise stages; I suppose I could put in an ODBC oracle stage but on this system ODBC is not configured and as a production system it would take a lot of paperwork to do.
Then I thought I could use user-defined SQL and just append a ";commit" to the line but that wouldn't parse correctly at runtime. Does this SQL allow multiple commands per line?
I wanted to run this question past this august forum to see if anyone might have encountered this type of scenario and might already have found a solution.
</a>