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

Home -> Community -> Usenet -> c.d.o.server -> Re: Index Insider

Re: Index Insider

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sat, 11 Nov 2000 14:15:01 -0000
Message-ID: <973948361.29278.0.nnrp-13.9e984b29@news.demon.co.uk>

I can't resist.

Let's assume that Oracle stores a depth
indicator on each branch (can't think why, it would be a terrible cost on branch block splits, but there is a hint that it does when you do an index treedump). But let's make the assumption that somewhere (either
stored, or in a piece of code) there is an explicit limiting.

The smallest space to store it would be
a signed byte - which would force a max
of 127 branch levels.

In the worst case, each branch points to just two blocks, so the number of leaf blocks would be power (2,127).

Since power(2,10) is a bit over power(10,3) let's approximate power(2,127) to
power(2,130) and call it power(10,3 * 13)

Somehow I don't think I'll try the experiment.

--

Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases

Publishers:                 Addison Wesley Longman
Book bound date:     8th Dec 2000
See a first review at:
http://www.ixora.com.au/resources/index.htm#practical_8i

Connor McDonald wrote in message <3A0BDDF6.10A4_at_yahoo.com>...


>I wonder what level you could actually achieve - take the smallest
>blocksize, largest index keys and keep on pumping rows in until you hit
>the table size limit then check the level...
>
>Anyone care to do the maths ??
>
Received on Sat Nov 11 2000 - 08:15:01 CST

Original text of this message

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