Hi neveraloneneverapart, welcome to the Monastery and to Perl, the One True Religion.

Assuming every line in __DATA__ is known to contain a number, I would use a hash to count results and a numeric comparison:

use strict; use warnings; use feature 'say';
use Data::Dumper;

my %results;

while (<DATA>) {
    chomp;
    $results{'zero'}++ if $_ == 0;
    $results{'pos'}++ if $_ > 0;
    $results('neg'}++ if $_ < 0;
}

print Dumper \%results;
[download]

Alternatively, to see if a string is a number, see Scalar::Util::looks_like_number().

If you really wanted to, for a regexp for Latin integers, you could use $num =~ m/  \A -? [0-9]+ \z /x;

To capture the sign (or nothing) and the digits and then be able to use them:

if ($num =~ m/  \A (-?) ([0-9]+) \z /x ) {
    my $sign = $1;
    my $digits = $2;
    ...
}
[download]

Also see Regexp::Common

Hope this helps!

edit: removed unnecessary escape slash