ksms.o
- From: <Satheesh.Babu@xxxxxxxxxxxxxxxxxx>
- To: <oracle-l@xxxxxxxxxxxxx>
- Date: Wed, 30 Jun 2004 16:15:48 +0530
Is there a way to open this file? Or atleast to view their contents.
Thanks and Regards,
Satheesh Babu.S
Associate Consultant.
080-57593938
Bangalore.
India.
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Lex de Haan
Sent: Wednesday, June 30, 2004 4:08 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: RE: Physics of the FILTER operation within SQL_PLANE.
well, the Oracle source code is rather complicated,=20
and I don't have access to it anymore ;-)
but there are *definitely* differences between RBO and CBO code paths;
RBO and CBO for sure start from different Oracle source code modules. =20
=20
Kind regards,
Lex.
=20
---------------------------------------------
visit my website at http://www.naturaljoin.nl=20
---------------------------------------------
=20
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of J.Velikanovs@xxxxxxxx
Sent: Wednesday, June 30, 2004 11:03
To: oracle-l@xxxxxxxxxxxxx
Subject: RE: Physics of the FILTER operation within SQL_PLANE.
Before this day I was sure that Optimizer code path working just to
create=20
Execution plan,=20
then feach, execute opeartions are executed exactly the same way=20
independently of used optimizator.
Please correct me if I am wrong.
Jurijs
"Lex de Haan" <lex.de.haan@xxxxxxxxxxxxxx>
Sent by: oracle-l-bounce@xxxxxxxxxxxxx
30.06.2004 13:00
Please respond to oracle-l
=20
To: <oracle-l@xxxxxxxxxxxxx>
cc: <oracle-l-bounce@xxxxxxxxxxxxx>
Subject: RE: Physics of the FILTER operation within=20
SQL_PLANE.
Jurijs,
I fully agree with you -- RBO is ideal for that purpose,
but there is a risk that you use RBO code path.
in my humble opinion (open for corrections!)
it is better to test things against the CBO, with "proper" statistics
available,
and then to force certain execution plans by using CBO hints...
(there are more than enough of them ;-)
Kind regards,
Lex.
---------------------------------------------
visit my website at http://www.naturaljoin.nl
---------------------------------------------
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of J.Velikanovs@xxxxxxxx
Sent: Wednesday, June 30, 2004 10:33
To: oracle-l@xxxxxxxxxxxxx
Cc: oracle-l@xxxxxxxxxxxxx; oracle-l-bounce@xxxxxxxxxxxxx
Subject: RE: Physics of the FILTER operation within SQL_PLANE.
Lex,
Thank for comment,
I use RBO just for proof of concept.
I am sure that the same situation can be emulated with CBO, but in this
case it was easy for me to use RBO.
>> about the optimizer not being able to calculate differences
Even if would use CBO it will not be able to calculate difference in
case
of FILTER operation, was my point.
Jurijs
"Lex de Haan" <lex.de.haan@xxxxxxxxxxxxxx>
Sent by: oracle-l-bounce@xxxxxxxxxxxxx
30.06.2004 12:30
Please respond to oracle-l
To: <oracle-l@xxxxxxxxxxxxx>
cc:
Subject: RE: Physics of the FILTER operation within
SQL_PLANE.
Hi Jurijs,
allow me one comment about your statement below,
about the optimizer not being able to calculate differences:
you are forcing RBO behavior, a technology now obsolete in 10g
and untouched by Oracle development for many, many years...
(interesting email thread, by the way)
Kind regards,
Lex.
---------------------------------------------
visit my website at http://www.naturaljoin.nl
---------------------------------------------
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of J.Velikanovs@xxxxxxxx
Sent: Wednesday, June 30, 2004 10:02
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: Physics of the FILTER operation within SQL_PLANE.
If your hypnotize is true (regerding remembering last value), then I
have
one more interesting conclusion:
The cost of the same SQL with the same execution plans can differ
depending on rows ordering in the driving rowset.
Like with index range scan and clustering factor, but with one little
difference - Oracle optimizer in case of FILTER operation can't evaluate
this cost (in case of range scan - clustering factor statistics
available).
Even more, if your hypnotize is true, then cost difference can appears
not
from bigger LIO count, but also because in one case Oracle need to
manage
hash table in other doesn't (just use last remembered value).
For my point demonstrating purpose, I have a little bit modified my
initial testcase (1,2).
Take a look: the same SQL, the same ExplPlan, time differs by 18% and
Oracle optimizer can't caltulate this difference.
Jurijs
TESTCACE 1
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D
SELECT /*+ RULE */ count(m.v) from main_tab m
where
exists
(select v from
filter_tab f where f.n=3Dm.n and f.v like 'a%')
TESTCACE 2
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D
SELECT /*+ RULE */ count(m.v) from main_tab m
where
exists
(select v from
filter_tab f where f.n=3Dm.n and f.v like 'a%')
-- Binary/unsupported file stripped by Ecartis --
-- Type: text/x-vcard
-- File: Lex de Haan.vcf
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
-- Binary/unsupported file stripped by Ecartis --
-- Type: text/x-vcard
-- File: Lex de Haan.vcf
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
-- Binary/unsupported file stripped by Ecartis --
-- Type: text/x-vcard
-- File: Lex de Haan.vcf
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
