Re: Slow Update

There are times when I get to feel really smug!

Maybe you could swap some details of the stats on that table, its indexes and the column in the critical predicates with Amit to help identify why the silly cardinality estimate has appeared.

Regards (and thanks)
Jonathan Lewis

On Fri, 3 Nov 2023 at 08:23, Holvoet Jo <J.Holvoet_at_dwl.be> wrote:

> I can find that exact same sql_id in our Oracle EBS db as well, and our
> execution plan is basically doing what Jonathan says : a NL join between
> VW_NSO_1 and a unique index scan on PO_LINE_LOCATIONS_U1 instead of your
> hash join with a full table scan. We end up with a cost of 6 for the update
> and execution times in the centisecond range (although here it is hardly
> ever executed – at most once every few days).
>
>
>
> Regards
>
> Jo
>
>
>
>
>

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