Re: flush shared_pool and query performance

(Repost because of network failure)

Don't be confused. Just trace the sql.
On the first hard parse after flushg shared pool, lots of recursive SQLs are executed either.
In your case, the overhead of parsing the recursive queries seems to have high cost.
On the third hard parse(4567), the recursive SQLs are already parsed so no overhead.

Yes, aboves are just guesswork. You know BAAG, don't you? Just show us your sql trace and tkprof report.

Dion Cho

PS) I can post a long query that consumes more that 10sec to be parsed. :)

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