Re: Serious error in int() function?

Path	csiph.com!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From	Gregory Ewing <greg.ewing@canterbury.ac.nz>
Newsgroups	comp.lang.python
Subject	Re: Serious error in int() function?
Date	Wed, 13 Apr 2016 23:51:26 +1200
Lines	36
Message-ID	<dn6q60FmpvfU1@mid.individual.net> (permalink)
References	<52f7516c-8601-4252-ab16-bc30c59c8306@googlegroups.com>
Mime-Version	1.0
Content-Type	text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding	7bit
X-Trace	individual.net p30CIu3bUwza5UBa0YsxGQUcs9TlvwwgfNcbHh+5GEuhpIhFEu
Cancel-Lock	sha1:kXuP8Dj5jRrBGPudkXwq7567X40=
User-Agent	Mozilla Thunderbird 1.0.5 (Macintosh/20050711)
X-Accept-Language	en-us, en
In-Reply-To	<52f7516c-8601-4252-ab16-bc30c59c8306@googlegroups.com>
Xref	csiph.com comp.lang.python:106935

Show key headers only | View raw

martin.spichty@gmail.com wrote:
> print int(float(2.8/0.1))
> 
> yields
> 
> 27
> 
> instead of 28!!

This is a consequence of the fact that the machine does
floating point arithmetic in binary, not decimal.

0.1 is not exactly representable as a binary floating
point number, and the result of the division comes out
very slightly less that 28:

 >>> 2.8/0.1
27.999999999999996

This is not unique to Python; you'll get the same
result from anything that uses the machine's native
floating point numbers.

When using floating point, you always need to be
prepared for slight inaccuracies. Usually you don't
notice them, because the results are rounded to some
reasonable number of decimal places before display.
But sometimes they show up, such as when using
int(), which truncates instead of rounding.

The best way to handle things like this depends on
the circumstances. If you tell us what you're trying
to achieve, we may be able to offer advice.

-- 
Greg

Thread

Serious error in int() function? martin.spichty@gmail.com - 2016-04-13 00:41 -0700
  Re: Serious error in int() function? Marko Rauhamaa <marko@pacujo.net> - 2016-04-13 10:55 +0300
  Re: Serious error in int() function? Alain Ketterlin <alain@universite-de-strasbourg.fr.invalid> - 2016-04-13 10:00 +0200
  Re: Serious error in int() function? Peter Otten <__peter__@web.de> - 2016-04-13 10:03 +0200
  Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-13 09:04 +0100
  Re: Serious error in int() function? Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2016-04-13 23:51 +1200
  Re: Serious error in int() function? "ast" <nomail@com.invalid> - 2016-04-14 08:52 +0200
    Re: Serious error in int() function? Christian Gollwitzer <auriocus@gmx.de> - 2016-04-14 09:25 +0200
    Accuracy of math.sqrt(), was Re: Serious error in int() function? Peter Otten <__peter__@web.de> - 2016-04-14 09:46 +0200
    Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-14 08:59 +0100
      Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-14 09:13 +0100
        Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-14 13:07 +0100
          Re: Serious error in int() function? Dennis Lee Bieber <wlfraed@ix.netcom.com> - 2016-04-14 20:36 -0400
            Re: Serious error in int() function? Christian Gollwitzer <auriocus@gmx.de> - 2016-04-15 04:38 +0200
              Re: Serious error in int() function? John Pote <johnhpote@o2.co.uk> - 2016-04-15 12:05 +0100
                Re: Serious error in int() function? alister <alister.ware@ntlworld.com> - 2016-04-15 11:50 +0000
            Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-15 09:41 +0100
              Re: Serious error in int() function? blindanagram@nowhere.net - 2016-04-15 09:41 +0100

csiph-web