Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: performance problem
Hi Martin,
Thanks for your suggestion. I am afraid I cannot guarantee that the column int_keyfld4 will always be unique. It is true for the current situation. I just mentioned it to illustrate the high cardinality of the non-unique index. The 'not exists' subquery should therefore retrieve a maximum of one record for each update. I fail to understand why Oracle needs to read 7000 blocks for this. And am I correct that your reformulated update statement will also work with a non-unique index (because of the antijoin)?
Geoffrey Received on Fri Jun 22 2001 - 04:04:12 CDT