Using hashes is one way to tackle the problem, as others have shown. But since your example suggests that the data is clumped - that is, numbers which are the same follow each other, you don't need a hash, as the program below shows.

I'm just showing it here for the sake of showing an alternative way. I don't expect it to be faster - hashes are pretty fast, and the sort is likely to dominate the running time anyway.

Abigail

#!/usr/bin/perl

use strict;
use warnings 'all';

my @info;

while (<DATA>) {
    my ($num, $info) = split ' ', $_, 2;

    if (@info && $num == $info [-1] [0]) {
        $info [-1] [2] ++;
    }
    else {
        push @info => [$num, $info, 1]
    }
}

print map {"@{$_}[0, 1]"} sort {$b -> [2] <=> $a -> [2]} @info;

__DATA__
1  info1
1  info1
1  info1
2  info2
3  info3
3  info3
4  info4
4  info4
4  info4
4  info4



$ ./count
4 info4
1 info1
3 info3
2 info2
$
[download]