Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question: 

#!c:/perl/bin/perl
$x=100.0;
$y=-100.01;
print $x+$y;

[download]
I am getting output as 
-0.0100000000000051

[download]
please help me out why this is coming ? how to ensure to get only -0.01

Replies are listed 'Best First'.
Re: can't Understand output
by moritz (Cardinal) on Apr 15, 2008 at 18:45 UTC
    Perl uses your CPU's floating point numbers under the hood which have some inaccuracies.

    You can use bigrat instead, or rely only on integers, or round after a few digits.

[reply]
Re: can't Understand output
by apl (Monsignor) on Apr 15, 2008 at 18:49 UTC
    Similar questions have been asked two or three times this year. For a good overview, please see When is a 2 not a 2?.
[reply]
Re: can't Understand output
by FunkyMonk (Bishop) on Apr 15, 2008 at 18:46 UTC
    Generally, computers store numbers with decimal places as floating point numbers, which aren't stored exactly. The Some solutions are

    • work with some multiple of the numbers (eg 100 for yours), so you could use $x = 10000; $y = -10001; print $x - $y;
    • use printf to round your answer, like so printf "%0.2f", $x+$y

[reply]
[d/l]
[select]
Re: can't Understand output
by tachyon-II (Chaplain) on Apr 15, 2008 at 23:52 UTC
[reply]
Re: can't Understand output
by casiano (Pilgrim) on Apr 16, 2008 at 07:58 UTC
    Yes, IEEE floating point numbers aren't exact representations. Use Math::BigFloat: 
    pp2@nereida:~/src/perl/testing$ cat -n addlong.pl
     1  use warnings;
     2  use strict;
     3  use Math::BigFloat ':constant';
     4
     5  my $x=100.0;
     6  my $y=-100.01;
     7  print (($x+$y)."\n");

    [download]
    And now: 
    pp2@nereida:~/src/perl/testing$ perl addlong.pl
-0.01

    [download]
    Best

    Casiano

[reply]
[d/l]
[select]
Re: can't Understand output
by swampyankee (Parson) on Apr 16, 2008 at 18:44 UTC

    Because 0.01 cannot be exactly represented as a finite sum of powers of two, just as 1/3 cannot be exactly represented as the finite sum of powers of ten. Floating point numbers cannot represent all rational, let alone all real, numbers in a given range.

    Information about American English usage here and here. Floating point issues? Please read this before posting. — emc

[reply]

Back to Seekers of Perl Wisdom