Since I just updated my previous suggestion I figured I would Benchmark things...
Not that it matters but I thought you might be interested.
Update There was an issue with how I ran the test before. Here are the updated code and results.
#!/usr/bin/perl -w
use strict;
use Benchmark;
my $count = 500000;
my $line = "1!2!3!4!5!6!7!8!9";
my @arr = split /!/, $line;
my @exclude = (2,5..8);
my @exclude_lkup;
$exclude_lkup[$_] = 1 for @exclude;
sub ikegami
{
my @tarr = @arr;
my @filtered = map $tarr[$_], grep !$exclude_lkup[$_], 0..$#tarr;
}
sub injun
{
my @tarr = @arr;
@tarr[@exclude] = ();
@tarr = grep $_, @tarr;
}
timethese (
$count,
{'Ikegami' => '&ikegami',
'InjunJoel' => '&injun'}
);
Results in
Benchmark: timing 500000 iterations of Ikegami, InjunJoel...
Ikegami: 16 wallclock secs (15.88 usr + 0.00 sys = 15.88 CPU) @ 314
+94.08/s (n=500000)
InjunJoel: 10 wallclock secs ( 11.05 usr + 0.00 sys = 11.05 CPU) @ 4
+5265.25/s (n=500000)
Though I'm not versed enough in O(n) notation to tell you why...
-InjunJoel
"I do not feel obliged to believe that the same God who endowed us with sense, reason and intellect has intended us to forego their use." -Galileo
|