Re: SELECT table.currval FROM dual faster way to do this???

Those stats certainly look highly suspect.

1. Check that DUAL still has only one row in it
2. Check the cache_size, and ORDER option for the sequence
3. Check the number of sequences you use compared to the init.ora values sequence_cache_entries (default 10) and sequence_cache_hash_entries (default 7)

BTW - Are you running Oracle Parallel Server ?

In most cases, a sequence call should do no disk read and only a couple of logical I/Os. It should certainly deliver only 1 row per execute.

--

Jonathan Lewis
Yet another Oracle-related web site: www.jlcomp.demon.co.uk

tim_mcconechy_at_my-deja.com wrote in message <7k5dlf$jlt$1_at_nnrp1.deja.com>...
>Is there a better way to do:
>SELECT pdmlogseq.currval FROM dual
>
>Seems like a lot of disk reads for a simple query?
>
>
>First load time: 1999-06-15/11:42:07
>Buffer gets: 3565 ratio 6
>Disk reads: 2102 ratio 3
>Rows delivered 26857 ratio 42
>Executions 641
>Parses 92
>Memory 18569
>Sorts 0
>Invalidations 0
>
>
>Sent via Deja.com http://www.deja.com/
>Share what you know. Learn what you don't.
Received on Tue Jun 15 1999 - 07:36:32 CDT