From: Joerg Bruehe
Date: October 25 2010 6:23pm
Subject: Re: Query precision issue
List-Archive: http://lists.mysql.com/mysql/223443
Message-Id: <4CC5CB38.8040803@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: quoted-printable

Sairam,


sorry to be so blunt, and please don't feel offended:

Sairam Krishnamurthy wrote:
> [[...]]
>=20
> Can somebody help me to find out the problem? I can very well truncate
> the trailing zeros when querying, but I am interested in finding why an=

> additional trailing zero returns an empty set.

You have not understood the concept of floating point data and of their
binary representation.
For a starter, please try to write the *exact* value 1.0/3.0 with a
finite number of decimal digits.

A basic paper on this was available here:
http://cch.loria.fr/documentation/IEEE754/ACM/goldberg.pdf

Any attempt to solve your problem by giving a different number of
trailing zeros is bound to fail. You best drop the approach of using
equality predicates for floating point data, it will not work in the
general case.


J=F6rg

--=20
Joerg Bruehe,  MySQL Build Team,  joerg.bruehe@stripped
               (+49 30) 417 01 487
ORACLE Deutschland B.V. & Co. KG,   Komturstrasse 18a,   D-12099 Berlin
Geschaeftsfuehrer: Juergen Kunz, Marcel v.d. Molen, Alexander v.d. Ven
Amtsgericht Muenchen: HRA 95603