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

Home -> Community -> Mailing Lists -> Oracle-L -> full-scan vs index for "small" tables

full-scan vs index for "small" tables

From: <oracle-l-bounce_at_freelists.org>
Date: Tue, 27 Jun 2006 08:03:33 -0500
Message-ID: <C970F08BBE1E164AA8063E01502A71CF5D5B0D@WIN02.hotsos.com>


On a slightly modified topic, for a long time I've had a problem with this section 13.5.1.2.3 of the Oracle documentation. It's one of those sections that was written 25 years ago and apparently never subjected to scientific scrutiny.

The part about "...which can be read in a single I/O call..." is one of those myths that makes sense when you hear it, but it's just not true. An index scan of a 1-row, 1-block table is more efficient than a full table scan of that table. Try it.

Performance of an Oracle database is NOT uniquely determined by how many OS read calls your application causes it to make.

Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com
Nullius in verba  

Hotsos Symposium 2007 / March 4-8 / Dallas Visit www.hotsos.com for curriculum and schedule details...

-----Original Message-----

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Ogun.Heper_at_turkcell.com.tr Sent: Tuesday, June 27, 2006 4:07 AM
To: sjaffarhussain_at_gmail.com; oracle-l_at_freelists.org Subject: RE: db file scattered reads

Hi,

First of all, the p3 parameter doesn't have to be equal to DB_FILE_MULTIBLOCK_READ_COUNT. The only explanation about this value in the documentation is;

P3 - The number of blocks (should be greater than 1)

Furthermore, documentation states that;

13.5.1.2.3 Small Table

If a table contains less than DB_FILE_MULTIBLOCK_READ_COUNT blocks under the high water mark, which can be read in a single I/O call, then a full table scan might be cheaper than an index range scan, regardless of the fraction of tables being accessed or indexes present.

So this can be your situation, if these are small lookup tables, then Oracle can decide to do a full table scan and wait on the db file scattered read event consequently, with a p3 value less than DB_FILE_MULTIBLOCK_READ_COUNT. For example in my test case this occurs for a small table which has got only two blocks.

PARSING IN CURSOR #2 len=25 dep=0 uid=46 oct=3 lid=46 tim=32288060382808 hv=468215388 ad='3382740c' select count(*) from t
END OF STMT
PARSE #2:c=0,e=132,p=0,cr=0,cu=0,mis=0,r=0,dep=0,og=4,tim=32288060382801 BINDS #2:

EXEC #2:c=0,e=78,p=0,cr=0,cu=0,mis=0,r=0,dep=0,og=4,tim=32288060383009
WAIT #2: nam='SQL*Net message to client' ela= 2 p1=1413697536 p2=1 p3=0
WAIT #2: nam='SQL*Net message from client' ela= 1479 p1=1413697536 p2=1 p3=0
WAIT #2: nam='db file sequential read' ela= 34 p1=25 p2=16009 p3=1
WAIT #2: nam='db file scattered read' ela= 60 p1=25 p2=16010 p3=2
WAIT #2: nam='SQL*Net message to client' ela= 2 p1=1413697536 p2=1 p3=0
FETCH #2:c=0,e=511,p=3,cr=4,cu=0,mis=0,r=1,dep=0,og=4,tim=32288060385189 WAIT #2: nam='SQL*Net message from client' ela= 13590 p1=1413697536 p2=1 p3=0

Actually, you need to investigate this further in order to fully understand the issue.

  1. For which segments this is occuring? Are they small?
  2. If there are a small number of segments hitting this event, why they are being aged out from the buffer cache so often? Can CACHE or BUFFER_POOL KEEP be an option?
  3. If these segments are not too small, why are they being full scanned?

Best regards.
Ogün Heper

-----Original Message-----

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Syed Jaffar Hussain Sent: Monday, June 26, 2006 6:21 PM
To: _oracle_L_list
Subject: db file scattered reads

Hi,

On one of our very busy OLTP Oracle 9i Rel. (9205) database on AIX 5.3 OS, 'db file scattered reads' wait event, most of the time, it shows 2, 1, 3 as p3 value.
What surprised me is that the value for db_file_multiblock_read_coutn is set to 8, but, db file scattered reads never oftenly shows 8 as p3 values while looking in v$session_wait view. There is another doubt about, 'buffer busy waits, reason code 130. On dizwell site,from the below link, the author explains about buffer busy waits explanation for reason code 130. When I look into metalink note for buffer busy waits reference note, the explanation is totally confused me and look totally different.
In metalink note, it says that the reason code 130 is like 'Block is being read by another session and no other suitable block image was found.. Whereas on dizwell has little bit, differenet explanation. Can anyone explain about this?

http://dizwell.com/main/index.php?option=com_jd-wiki&Itemid=112&id=bufferbusywaits
--

Best Regards,
Syed Jaffar Hussain
8i,9i & 10g OCP DBA
Banque Saudi Fransi,
Saudi Arabia

I blog at :http://jaffardba.blogspot.com/ http://www.oracle.com/technology/community/oracle_ace/ace1.html#hussain



"Winners don't do different things. They do things differently."
--

http://www.freelists.org/webpage/oracle-l

Bu elektronik posta ve onunla iletilen bütün dosyalar sadece göndericisi tarafindan almasi amaclanan yetkili gercek ya da tüzel kisinin kullanimi icindir. Eger söz konusu yetkili alici degilseniz bu elektronik postanin icerigini aciklamaniz, kopyalamaniz, yönlendirmeniz ve kullanmaniz kesinlikle yasaktir ve bu elektronik postayi derhal silmeniz gerekmektedir. TURKCELL bu mesajin icerdigi bilgilerin doğruluğu veya eksiksiz oldugu konusunda herhangi bir garanti vermemektedir. Bu nedenle bu bilgilerin ne sekilde olursa olsun iceriginden, iletilmesinden, alinmasindan ve saklanmasindan sorumlu degildir. Bu mesajdaki görüsler yalnizca gönderen kisiye aittir ve TURKCELLin görüslerini yansitmayabilir Bu e-posta bilinen bütün bilgisayar virüslerine karsi taranmistir.

This e-mail and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you are not the intended recipient you are hereby notified that any dissemination, forwarding, copying or use of any of the information is strictly prohibited, and the e-mail should immediately be deleted. TURKCELL makes no warranty as to the accuracy or completeness of any information contained in this message and hereby excludes any liability of any kind for the information contained therein or for the information transmission, reception, storage or use of such in any way whatsoever. The opinions expressed in this message belong to sender alone and may not necessarily reflect the opinions of TURKCELL. This e-mail has been scanned for all known computer viruses.

--

http://www.freelists.org/webpage/oracle-l

--

http://www.freelists.org/webpage/oracle-l Received on Tue Jun 27 2006 - 08:03:33 CDT

Original text of this message

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