Re: substr or like

From: <chmanu_at_gmail.com>
Date: Thu, 3 Jul 2008 09:44:45 -0700 (PDT)
Message-ID: <0568b1bf-7a0b-4376-8438-9803cc58f53d@i76g2000hsf.googlegroups.com>

> Run the built-in SQL trace facility on the two SQL statements or use
> auto-trace.  Remember the effect that having some of the data buffered
> from a prior run will have on subsequent runs so if possible pre-load
> the data or take an average of the followup executions of the SQL to
> base you analysis on.
>
> Parse information is available via the trace.  See Performance and
> Tuning manual for instructions on use.
>
> HTH -- Mark D Powell --

Thanks for your responses, i have launched the statistic script cause they re was not up todate and now, the two queries take the same time : 14 second.

All my tests are done many times and i take those when pre load is ok.

The most cost is a table full scan on the substr query which increase the time but without be more than with like operator query (using a index).

I will use a substr index and it will be better.

Chmanu Received on Thu Jul 03 2008 - 11:44:45 CDT

Original text of this message