Event 10053 and flushing
March 5, 2007 Leave a comment
At various points today, I have convinced myself that various daft things were happening with Oracle, my hints and/or SQL*Plus when the real cause was just me being really daft.
I’ve been looking at a couple of problem queries today and using extended tracing of the CBO using event 10053.
However, every now and then my tracing would cease to provide anything meaningful.
The reason was that the CBO wasn’t doing it’s calculations because I was repeatedly running the same queries and much of the time the execution plans were in the shared pool.
So I should have been flushing the shared pool, changing the string literals in the SQL, or adding a space here or there to force a hard parse.