Groups > comp.lang.python > #20691

Re: Python math is off by .000000000000045

From	Christian Heimes <lists@cheimes.de>
Subject	Re: Python math is off by .000000000000045
Date	2012-02-22 19:26 +0100
References	<CAO+9iGcg5aTuv1y_GYEMJAPw8zhOmS=tDKTjVsaBe77Q5cd6TA@mail.gmail.com>
Newsgroups	comp.lang.python
Message-ID	<mailman.51.1329935200.3037.python-list@python.org> (permalink)

Show all headers | View raw

Am 22.02.2012 19:13, schrieb Alec Taylor:
> Simple mathematical problem, + and - only:
> 
>>>> 1800.00-1041.00-555.74+530.74-794.95
> -60.950000000000045
> 
> That's wrong.

That's only the correct answer for unlimited precision, not for IEEE-754
semantics. http://en.wikipedia.org/wiki/IEEE_754

> Proof
> http://www.wolframalpha.com/input/?i=1800.00-1041.00-555.74%2B530.74-794.95
> -60.95 aka (-(1219/20))
> 
> Is there a reason Python math is only approximated? - Or is this a bug?

Python uses the platforms double precision float datatype. Floats are
almost never exact.

Christian

Back to comp.lang.python | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

Re: Python math is off by .000000000000045 Christian Heimes <lists@cheimes.de> - 2012-02-22 19:26 +0100
  Re: Python math is off by .000000000000045 Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2012-02-22 22:21 +0000

csiph-web