Yes, you are right. The plan I got from sql_plan is different from the one
I got via explain plan.
thank you
Gene Gurevich
Oracle Engineering
224-405-4079
"Ghassan Salem"
<salem.ghassan@(protected)
ail.com> To
Sent by: genegurevich@(protected)
oracle-l-bounce@(protected)
reelists.org oracle-l <oracle-l@(protected)>
Subject
Re: cursor sharing
06/20/2006 12:31
PM
Please respond to
salem.ghassan@(protected)
il.com
Well,
you better get the plan from v$sql_plan, as the explain plan does not do
bind-peeking (when you use the cursor_sharing=force setting). so the plan
given by explain can be different from what you have when you actually
execute the statement (depeding on bind values, whether this is the first
time you execute the statement or not , and the presence of histograms)
rgds
On 6/20/06, genegurevich@(protected) <
genegurevich@(protected):
No. I am using the following to get the explain plan:
select plan_table_output from
table(dbms_xplan.display('PLAN_TABLE',null,'ALL')) (found it on metalink)
thank you
Gene Gurevich
Oracle Engineering
224-405-4079
"goran
bogdanovic"
<goran00@(protected)
To
m>
genegurevich@(protected)
cc
06/20/2006 04:20 oracle-l <oracle-l@(protected) >
AM
Subject
Re: cursor sharing
Did you checked in v$sql_plan the real execution plan with
CURSOR_SHARING_EXACT ?
genegurevich@(protected):
Hi all:
Can someone recommend a good article on the cursor sharing in
Oracle
9. I
just found a query that runs in 5 seconds
with CURSOR_SHARING_EXACT and in more than 1 hour without it
(cursor_sharing is set to force on the DB level)
I'm trying to find out why it happens.
thank you
Gene Gurevich
--
http://www.freelists.org/webpage/oracle-l
--
http://www.freelists.org/webpage/oracle-l
--
http://www.freelists.org/webpage/oracle-l