Here's my attempt at an answer. Personally speaking, I love hashes. You just have to be aware of the limitations of which approach you take (array vs. hash).

I created a sub "isNull" as a homegrown replacement for "exists". The idea is to check your assumptions (and data) before you operate on that data.

#!/usr/bin/perl -w

use strict;
sub keyExists;

# Begin Main
my %hash1 = ();  # Initialize empty hash.  Perhaps unnecessary?
%hash1 = (
    "1", "20",
    "2", "20",
    "4", "19",
    "5", "20",
    "10", "20",
    "6", "18");
my %hash2 = ();  # Initialize empty hash.  Perhaps unnecessary?
%hash2 = (
    "1", "19",
    "2", "20",
    "4", "16",
    "5", "19",
    "6", "20");

foreach my $thisKey (sort keys %hash1) {
    # Check whether key from hash2 returns null to protect from undefi
+ned key for hash2
    if (isNull(\%hash2, $thisKey)) {
        print "hash2 contains null value for key: $thisKey", "\n"; 
    }
    else {
        my $result = $hash1{$thisKey} - $hash2{$thisKey};
        print "$thisKey $hash1{$thisKey} minus $hash2{$thisKey} equals
+ $result", "\n";
    }
} # End foreach

exit 0;  # End Main

sub isNull {
    my $hashref = shift;    # Reference to hash
    my $key = shift;    # key to check
    my $rc = 0;
    
    if ($hashref->{$key}) {
        $rc = 0;    # Hash returns non-null value
    }
    else {
        $rc = 1;    # Hash returns Null value
    }
    return $rc;
}  # End sub isNull
[download]