#!/usr/bin/env perl
# https://www.perlmonks.org/?node_id=11148465

use warnings;
use strict;
use feature qw{ say };
use Scalar::Util qw{ dualvar };
use MCE;

die "Usage: $0 input1 [ input2 ... ]\n" unless @ARGV;

# Ensure given input files are readable.
my @infiles = @ARGV; @ARGV = ();

for (@infiles) {
    die "Cannot open '$_'" unless -r "$_";
}

# MCE gather and parallel routines.
our @DATA;

sub gather_routine {
    my ($data_ref) = @_;
    while (@{ $data_ref }) {
        push @DATA, dualvar(
            shift @{ $data_ref },
            shift @{ $data_ref }
        );
    }
}

sub parallel_routine {
    my ($char, %by_word, @data, @ret) = ($_);

    for my $file (@infiles) {
        open my $fh, '<', $file;
        while (<$fh>) {
            if (substr($_,0,1) eq $char) {
                chomp;
                my ($k, $v) = split /\t/, $_;
                $by_word{$k} += $v;
            }
        }
        close $fh;
    }

    while (my ($k, $v) = each %by_word) {
        push @data, dualvar($v, $k);
    }

    push(@ret, 0+$_, "$_") for sort @data;

    MCE::relay { MCE->gather(\@ret) };
}

# Run parallel using MCE.
warn "start\n";
my $tstart1 = time;

MCE->new(
    input_data  => ['a'..'z'],
    max_workers => 7,
    chunk_size  => 1,
    init_relay  => 1,
    posix_exit  => 1,
    gather      => \&gather_routine,
    user_func   => \&parallel_routine,
    use_threads => 0,
)->run(1);

my $tend1 = time;
warn "get properties + pre-sort: ", $tend1 - $tstart1, " secs\n";

# Output dualvar data, sorted by count.
$| = 0;  # enable output buffering

my $tstart2 = time;
say "$_\t".(0+$_) for sort { $b <=> $a } @DATA;
my $tend2 = time;

warn "final sort + output: ", $tend2 - $tstart2, " secs\n";
warn "total: ", $tend2 - $tstart1, " secs\n";
[download]