Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Query Performance: Using DECODE() instead of IN()

Re: Query Performance: Using DECODE() instead of IN()

From: Job Miller <jobmiller_at_yahoo.com>
Date: 20 Oct 2003 14:09:39 -0700
Message-ID: <2edf8ed9.0310201309.7b718a5d@posting.google.com>


use sql_trace/tkprof, autotrace, execution plans with real volumes of data to see how these perform.

actually runstats is a great way also to compare two things like this also.

http://asktom.oracle.com/~tkyte/runstats.html

autotrace shows the opposite of what you propose below

>
> I think I recall somewhere that using a decode uses an index scan where
> using IN() uses a full table scan, however I can't seem to prove this.
>
> Trying to search for in and decode in google is not fun. Any ideas?
>
> PT
Received on Mon Oct 20 2003 - 16:09:39 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US