bignum frustration

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

Hey, all. What'd I goof up here?

use bignum;

$foo = 12345.6789;
$bar = $foo->copy()->bfround(0);
$baz = $foo - $bar;
print "what's wrong here? $baz\n";

$byhand = 12345.6789 - 12345;
print "as expected $byhand\n";
[download]

... which does this:

what's wrong here? 0
as expected 0.6789
[download]

So the subtraction is getting zero, where I expected roughly 0.6789 also. And yes, I get the same thing if i use $baz = $foo->bsub($bar).
Any hints appreciated. Thanks!

Update - Thanks for the welcome; I've been lurking and annonning for years so it was prolly time. It was indeed a contamination of precision. When doing the bfround(0) you can Dumper the result and see it picked up a p=0 member, which got carried along after that. The bfround(0)->bfround(-5) solution works but I need to rtfm a couple more times to get my head around it. The problem I was trying to solve was merely, "what's the distance from this float to the nearest integer, wherever that is"

Comment on bignum frustration Select or Download Code

Replies are listed 'Best First'.
Re: bignum frustration by SuicideJunkie (Vicar) on Aug 15, 2012 at 21:39 UTC
`use bignum; $foo = 12345.6789; $bar = $foo->copy()->bfround(0)->bfround(-5); $baz = $foo - $bar; print "what's wrong here? $foo - $bar = $baz\n"; $byhand = 12345.6789 - 12345; print "expected $byhand\n";` [download] Gives: `what's wrong here? 12345.6789 - 12346.00000 = -0.32110 expected 0.6789` [download] Seems to me that the precision is propagated to the result value, and -0.32 rounds up to zero in your original example.	[reply] [d/l] [select]
Re: bignum frustration by Athanasius (Archbishop) on Aug 16, 2012 at 03:26 UTC
Hello not_japh, and welcome to the Monastery! `#! perl use strict; use warnings; use bignum; my $foo = 1_234_500_000_000_000_000_000_000.6789; my $bar = int $foo; my $baz = $foo - $bar; print 'baz = ', $baz, "\n";` [download] Output: `baz = 0.6789` [download] Often, the simplest solution is also the best. `;-)` Update: Added greeting. Athanasius <°(((>< contra mundum	[reply] [d/l] [select]
Re^2: bignum frustration by Anonymous Monk on Aug 16, 2012 at 14:38 UTC
How can you be >0.5 away from the nearest integer?	[reply]
Re^3: bignum frustration by Athanasius (Archbishop) on Aug 16, 2012 at 15:39 UTC
Finding the nearest integer is still easy with just the `int` function: `#! perl use strict; use warnings; use bignum; while (my $num = <DATA>) { chomp $num; print 'float = ', $num, ', nearest int = ', int ($num + 0.5), "\n" +; } __DATA__ 12345.6789 1_234_500_000_000_000_000_000_000.6789 10.49 11.5 12.500000001 1_234_500_000_000_000_000_000_042.499999 13` [download] Output: `float = 12345.6789, nearest int = 12346 float = 1_234_500_000_000_000_000_000_000.6789, nearest int = 12345000 +00000000000000001 float = 10.49, nearest int = 10 float = 11.5, nearest int = 12 float = 12.500000001, nearest int = 13 float = 1_234_500_000_000_000_000_000_042.499999, nearest int = 123450 +0000000000000000042 float = 13, nearest int = 13` [download] `:-)` Athanasius <°(((>< contra mundum	[reply] [d/l] [select]