Re: sequential read on full-table scan?

From: Paul Baumgartel <paul.baumgartel_at_gmail.com>
Date: Fri, 20 May 2005 17:34:13 -0400
Message-ID: <f8c477105052014347376c75e@mail.gmail.com>

Zoran,

I took your advice and found some interesting information. Here's a typical snippet of the trace file:

WAIT #64: nam=3D'db file scattered read' ela=3D 1150 p1=3D17 p2=3D9813 p3= =3D3
WAIT #64: nam=3D'db file scattered read' ela=3D 959 p1=3D17 p2=3D9817 p3=3D= 2
WAIT #64: nam=3D'db file sequential read' ela=3D 2040 p1=3D17 p2=3D9820 p3= =3D1
WAIT #64: nam=3D'db file sequential read' ela=3D 547 p1=3D17 p2=3D9822 p3= =3D1

Note that there ARE scattered reads going on as well as sequential, but the scattered reads are grabbing only 2 or 3 blocks at a time. DB file multiblock read count is set to 16.

Paul

On 5/20/05, Martic Zoran <zoran_martic_at_yahoo.com> wrote:

> Paul,

>=20

> Wanted to say fact 3 (buffer in the cache) + sometimes
> fact 2 (extent edge).

>=20

> Zoran

>=20
>=20
>=20
>=20
>=20

> __________________________________
> Do you Yahoo!?
> Make Yahoo! your home page

> http://www.yahoo.com/r/hs
>=20

--=20
Paul Baumgartel
paul.baumgartel_at_gmail.com

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

Received on Fri May 20 2005 - 17:38:54 CDT