Hi,
I've run across a situation where a series of updates are being done on a
table. If the update succeeds, it takes around 2 ms. If it fails (as it
frequently does), it takes around 200 ms. The where portion of the update
is a perfect match for the unique primary index. An explain indicates that
it is accessing this index. Analysis of the index confirms that it is
unique. Nothing glaring in the index or table statistics. If it is using
a unique index, why would an unsuccessful search take any longer than a
successful search?
Thanks,
Russ Brooks
Received on Sat Jul 11 1998 - 12:53:25 CDT
