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 -> Parsing in cost-based mode on Oracle 7.3

Parsing in cost-based mode on Oracle 7.3

From: Michel Tilman <mctilman_at_cargo.be>
Date: 1998/02/18
Message-ID: <6cfhbh$39c4@afrodite.telenet-ops.be>#1/1

Hello,

We recently switched to cost-based optimizer. While on the average the resulting parsed and cached queries give good access time, the actual parsing process takes a lot of time in several cases, e.g. nearly 3 min parsing (first time) vs 4 sec actual execution (involving say 7 tables). As we have a rather dynamic query interface, whether a query (plan) will remain cached is not always very predictable. Can someone give me some insight in what goes on in the actual analysis phase (query parsing) to understand what is going on? Or am I missing something? As far I know our statistics are updated weekly, and the population grows rather linearly. We use parameter binding. We have several hundred tables, the largest being about 100.000 record (just a few, the others are much smaller).

Thanks in advance,

Michel Tilman
Argo Received on Wed Feb 18 1998 - 00:00:00 CST

Original text of this message

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