use v5.12;
use warnings;
use Test::More;
use Benchmark qw/cmpthese/;


my $AoH;
my $density = 75;

for my $n_rec (1, 10, 100, 1000) {
    say "";
    say "=== num of records is: ", $n_rec, " duplication: $density%";
    $AoH = create_data ($density, $n_rec);
    is_deeply (
        [ sort &list_join  ],
        [ sort &slice_join ],
        );
    is_deeply (
        [ sort &grep_map   ],
        [ sort &slice_join ],
        );
    is_deeply (
        [ sort &map_map    ],
        [ sort &slice_join ],
        );

    cmpthese ( -1, {
        map_map    => \&map_map,
        grep_map   => \&grep_map,
        list_join  => \&list_join,
        slice_join => \&slice_join,
        });

    }

done_testing;

sub list_join {
    my %H;
    %H = (%H, %$_) for @$AoH;
    return keys %H;
    } # list_join

sub slice_join {
    my %H;
    @H{keys %$_} = () for @$AoH;
    return keys %H;
    } # slice_join

sub grep_map {
    my %seen;
    return grep { !$seen{$_}++ } map { keys %$_ } @$AoH;
    } # grep_map

sub map_map {
    my %H = map { $_ => 1 } map { keys %$_ } @$AoH;
    return keys %H;
    } # map_map

sub create_data {
    my ($density, $records) = @_;

    my @AoH;
    push @AoH, {map { rand 100 <= $density ? ( $_ => $_ ) : () } "A" .
+. "ZZ"}
        for 1 .. $records;
    return \@AoH;
    }
[download]

=== num of records is: 1 duplication: 75%
ok 1
ok 2
ok 3
              Rate  list_join    map_map   grep_map slice_join
list_join   4575/s         --       -26%       -40%       -59%
map_map     6224/s        36%         --       -18%       -44%
grep_map    7625/s        67%        23%         --       -31%
slice_join 11079/s       142%        78%        45%         --

=== num of records is: 10 duplication: 75%
ok 4
ok 5
ok 6
             Rate  list_join    map_map   grep_map slice_join
list_join   202/s         --       -71%       -77%       -87%
map_map     696/s       244%         --       -21%       -57%
grep_map    884/s       337%        27%         --       -45%
slice_join 1612/s       697%       131%        82%         --

=== num of records is: 100 duplication: 75%
ok 7
ok 8
ok 9
             Rate  list_join    map_map   grep_map slice_join
list_join  17.5/s         --       -72%       -84%       -89%
map_map    63.1/s       260%         --       -43%       -61%
grep_map    110/s       526%        74%         --       -32%
slice_join  161/s       821%       156%        47%         --

=== num of records is: 1000 duplication: 75%
ok 10
ok 11
ok 12
            (warning: too few iterations for a reliable count)
             Rate  list_join    map_map   grep_map slice_join
list_join  1.73/s         --       -73%       -85%       -89%
map_map    6.40/s       270%         --       -46%       -60%
grep_map   11.8/s       583%        85%         --       -26%
slice_join 16.0/s       825%       150%        35%         --
1..12
[download]