RE: Deletion Of 160 Million Rows.

From: david wendelken <davewendelken_at_earthlink.net>
Date: Tue, 8 Feb 2005 13:48:30 -0800 (PST)
Message-ID: <10755172.1107899310586.JavaMail.root@beaker.psp.pas.earthlink.net>

Aw, c'mon Ryan, tell us how you really feel about it.

So, if I add row-level delete triggers, can I have a block AND a row-based process?

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

From: ryan_gaffuri_at_comcast.net
Sent: Feb 8, 2005 1:35 PM
To: thomas.mercadante_at_labor.state.ny.us,

"'kristian_at_kmja.com'" <kristian_at_kmja.com>, oracle-l_at_freelists.org Cc: "Mercadante, Thomas F" <thomas.mercadante_at_labor.state.ny.us> Subject: RE: Deletion Of 160 Million Rows.

I will say this one more time and speaking from experience.... I have worked with large databases in the multi-multi-terabyte range. The other ideas here are wrong. do not follow them.

The number of records involved is FAR less important than the volume of data. No matter what shop I go to, it takes the longest time to get this concept across to people. Be they DBAs or whatever. When doing bulk processing you are reading the whole table. It is a 'block' algorithm, not a row algorithm.

--

http://www.freelists.org/webpage/oracle-l Received on Tue Feb 08 2005 - 16:51:22 CST