benchmarking uniq

bench.pl

#!/usr/bin/env perl
use 5.010;
use strict;
use warnings;

use Benchmark qw(:hireswallclock cmpthese);
use List::MoreUtils qw(uniq);

my @data_sizes = (15_000, 150_000);
# my @data_sizes = (5_000, 20_000);
my $key_count = 10;
my $iters = 100;

sub build_stuff {
	my ($size) = @_;
	my @stuff;
	push @stuff, { map { $_ => 1 } 1..$key_count } for 1..$size;
	return @stuff;
}

for my $size (@data_sizes) {
	my @stuff = build_stuff($size);
	cmpthese($iters, {
		"scoped $size uniq"   => sub { uniq map { keys %$_ } @stuff },
		"unscoped $size uniq" => sub { uniq map keys %$_, @stuff },
		"map in hash slice"   => sub { my %uniq; @uniq{map keys %$_, @stuff} = () },
	});
}