G'day hiyesthanks,

Welcome to the Monastery.

Please post your results with your question. Guidelines for posting questions can be found in "How do I post a question effectively?". You can update your post with the results; "How do I change/delete my post?" explains how to do this.

A screenshot of your terminal is an inappropriate way to present your results; however, when you do have a link to post, please create an actual link (not a URL as plain text) — "What shortcuts can I use for linking to other information?" has details about this.

From the code you posted, I'd say the first thing you should do is read "perlretut - Perl regular expressions tutorial".

Here's the regexes you could have used to match the numbers in your DATA section:

$ perl -E '/^0$/ and say for qw{19 -22 498 512 15 -932 0 22 808 17 -32
+}'
0
$ perl -E '/^[1-9]\d*$/ and say for qw{19 -22 498 512 15 -932 0 22 808
+ 17 -32}'
19
498
512
15
22
808
17
$ perl -E '/^-[1-9]\d*$/ and say for qw{19 -22 498 512 15 -932 0 22 80
+8 17 -32}'
-22
-932
-32
[download]

That was run on "macOS 10.12.5" using "Perl 5.26.0". I don't have MSWin available to run a comparison; although, I'd be highly surprised if the results differed.

If you believe you're getting different results on various platforms, you should provide details of the OS and Perl you used (as I did above).

This isn't "perfect", but gives an idea. I would use better variable names so that you have a chance to understand them 5 years from now.

#!usr/bin/perl

# The program calculates the total zeros and positive integers from 
# the data using regex

use strict;
use warnings;
my ( $c_positive_int, $c_negative_int, $c_zero_int, $ctr_not_int) = ( 
+0, 0, 0, 0 );

while( my $num = <DATA> ) 
{
        $num =~ s/^\s+//;  # delete leading space(s)
        $num =~ s/\s+$//;  # delete trailing space(s)
        
        if ( $num =~ /^0$/ )      # integer single 0
        {
                $c_zero_int++;
        }
        elsif ( $num =~ /^\d+$/ ) # only positive digits
        {
                $c_positive_int++;
        }
        elsif ( $num =~ /^-\d+$/ ) #leading minus sign, then only digi
+ts 
        {
                $c_negative_int++;
        }
        else
        {
            $ctr_not_int++;    #non integer 
            print "**** Error $num is not an integer!\n";
        }
}
printf "freq(Z+):%8i\n", $c_positive_int;
printf "freq(Z-):%8i\n", $c_negative_int;
printf "freq(0): %8i\n", $c_zero_int;
printf "freq(?): %8i\n", $ctr_not_int;
printf "Total:   %8i\n", $c_positive_int + $c_negative_int + 
                         $c_zero_int + $ctr_not_int ;

=Results
**** Error 32.5 is not an integer!
freq(Z+):       7
freq(Z-):       3
freq(0):        1
freq(?):        1
Total:         12

=cut

__DATA__
19
-22
498
512
15
-932
0
22
808
17
-32 
32.5
[download]

You've already gotten good answers regarding CRLF, and chomping, and alternate regular expressions. However, to me, regex doesn't seem the best tool for the job. Testing whether a number is negative, zero, or positive seems a job for numerical operators to me. Using Test::More to verify and Benchmark to compare times, with the same set of 1000 random random integers:

__END__
1..4
# grep => {
#   'count_of_negatives' => 479,
#   'count_of_others' => 0,
#   'count_of_positives' => 506,
#   'count_of_zeros' => 15,
#   'total_count' => 1000
# }
# comparison_ops => {
#   'count_of_negatives' => 479,
#   'count_of_others' => 0,
#   'count_of_positives' => 506,
#   'count_of_zeros' => 15,
#   'total_count' => 1000
# }
# spaceship_op => {
#   'count_of_negatives' => 479,
#   'count_of_others' => 0,
#   'count_of_positives' => 506,
#   'count_of_zeros' => 15,
#   'total_count' => 1000
# }
# [hiyesthanks] regex => {
#   'count_of_negatives' => 521,
#   'count_of_others' => 0,
#   'count_of_positives' => 464,
#   'count_of_zeros' => 15,
#   'total_count' => 1000
# }
# [Marshall] regex => {
#   'count_of_negatives' => 479,
#   'count_of_others' => 0,
#   'count_of_positives' => 506,
#   'count_of_zeros' => 15,
#   'total_count' => 1000
# }
ok 1 - verify comparison operators give same results as grep
ok 2 - verify spaceship operators give same results as grep
not ok 3 - verify [hiyesthanks] regex give same results as grep
#   Failed test 'verify [hiyesthanks] regex give same results as grep'
#   at C:\usr\local\share\passthru\perl\perlmonks\1201647.pl line 22.
#     Structures begin differing at:
#          $got->{count_of_positives} = '464'
#     $expected->{count_of_positives} = '506'
ok 4 - verify [Marshall] regex give same results as grep
                         Rate use_marshall_regex use_hiyesthanks_regex
+ use_grep use_spaceship_op use_comparison_ops
use_marshall_regex     1715/s                 --                  -18%
+     -26%             -33%               -84%
use_hiyesthanks_regex  2088/s                22%                    --
+      -9%             -18%               -81%
use_grep               2307/s                35%                   10%
+       --              -9%               -79%
use_spaceship_op       2543/s                48%                   22%
+      10%               --               -77%
use_comparison_ops    10857/s               533%                  420%
+     371%             327%                 --
# Looks like you failed 1 test of 4.
[download]

edit: fixed bugs, which were verifying and printing the wrong versions; benchmark was ok

Dave.

Just to confirm what the others have said, if you save your file with CRLF line endings on a system that uses LF line endings (*NIX, including Mac OS X) and run it, it gives the wrong results. If you then put binmode DATA, ':crlf'; just before the loop, it works again. :crlf is the default on Windows, see open and binmode, and it causes the CRLF line endings to be converted to LF, so that chomp works properly. Without :crlf, what is happening is that $num contains a trailing CRLF, but chomp only removes the LF, leaving the strings looking like "19\r", which you can see yourself if you use Data::Dump or Data::Dumper with $Data::Dumper::Useqq=1; - see also the Basic debugging checklist. The best solution IMO is to convert the file to LF line endings, for example using fromdos from the Tofrodos package, and also many text editors support selecting the line endings in the "Save As" dialog.

Also, regarding your regexes, I just wanted to point out Regexp::Common::number, which you might find useful, and also that the regex /^[0].{0}/ looks a little suspicious: .{0} doesn't really do anything, and rewriting the regex as /^0/ shows that it will accept any string beginning with a 0, including stuff like "0 but true", is that really what you want?