RE: Mutating (SQL) execution plan!?...is that possible...strangest thing I have every seen

Hi Chris

>If developer changes something trivial, like the case of the word and,
>like; "AND" to "and", in he Perl code and runs it again and it takes
>only seconds!!!
>Now I realize that maybe Oracle sees this as technically new SQL, but =
it
>is the same query and they should use the same plan!?
>Finally as soon as we flush the shared pool the original query (in the
>program code) starts performing within seconds again!?

Did you check the execution plan associated with the cursor in the = shared pool with one of the V$SQL_PLAN* views?

Chris

--
http://www.freelists.org/webpage/oracle-l