Re^2: sorting HoH according to value

Definitely less compact, arguably more readable — although I tend to read the compact stuff more easily.

But which is faster? I know I was expecting a speedup from using each instead of keying each by hand. But does the sub slow it down?

use strict;
use Benchmark qw(cmpthese);

my %bdry;
for my $i ( 1 .. 35 ) {
    for my $j ( 1 .. 35 ) {
        $bdry{$i}{$j} = 1+ int rand 50;
    }
}

cmpthese(1_000, { jetsway=>\&jetsway, mwahsway=>\&mwahsway });

sub jetsway {
    my @keys =
        sort { $bdry{ $b->[0] }{ $b->[1] } <=> $bdry{ $a->[0] }{ $a->[
+1] } }
        map { my $k = $_; map { [$k, $_] } keys %{ $bdry{$k} } }
        keys %bdry;

    return @keys;
}

sub mwahsway {
    my @flipped_hash = reverse sort {$a->[0] <=> $b->[0] } flippout (%
+bdry);

    return @flipped_hash;
}

sub flippout {
    my %h = @_;
    my @f;
    while( my ($k,$v) = each %h ) {
        while( my ($vk,$vv) = each %$v) {
            push @f, [$vv, $k, $vk]
        }
    }
    return @f
}
[download]

          Rate  jetsway mwahsway
jetsway  103/s       --     -47%
mwahsway 195/s      89%       --

-Paul

Comment on Re^2: sorting HoH according to value Download Code