I use this regularly to compare columns in tab-delimited files. It could be quite easily modified to suit your purposes.

I wouldn't normally just hand you code, but I had this already. If you want the results to go to a file instead of a screen, all you need to do it compcolumns.pl file1 column1 file2 column2 > outputfile.txt:

#! perl -w
use strict;
my ($file, $col, $file2, $col2) = @ARGV;
my (%unique, %unique2);
my @column;
my ($x, $header);

my @fields;
# I put something like this in all my scripts so I don't need to remem
+ber argument order
if ($file eq "?"){
    die "usage:  compcolumns.pl file1 column1 file2 column2\n";
    }
    

## whenever you open a file, do yourself a favor and tell yourself if 
+it doesn't open
open FIL, $file or die "Could not open $file: $!\n";
$x = 0;
%unique=();
while (<FIL>){
    chomp;
        ##The split here, a regular expression, is what you would chan
+ge (at least if your data is based on space delimiting instead of tab
+-delimiting).
    @fields = split /\t/;
    
    if (defined ($fields[$col]) ){
        $header = $fields[$col] if $x == 0;
        $unique{$fields[$col]}++;
            }
    else{
        $unique{"BLANK"}++;
        $header = "BLANK";
        }
    $x++;    
    }
    
open FIL, $file2 or die "Could not open $file2: $!\n";
$x=0;
while (<FIL>){
    chomp;
    @fields = split /\t/;
    if (defined ($fields[$col2]) ){
            $header = $fields[$col2] if $x == 0;
            $unique2{$fields[$col2]}++;
                }
        else{
            $unique2{"BLANK"}++;
            $header = "BLANK";
            }
    $x++;
    }
## The output routine.  Note that if something's in both files, it wil
+l be printed twice.      
foreach (sort keys %unique) {
    if (!exists $unique2{$_}){
        print "$_ in $file but not $file2\n";
        }
    else {
        print "$_ in both files\n";
        }
    }

foreach (sort keys %unique2) {
    if (!exists $unique{$_}){
        print "$_ in $file2 but not $file\n";
        }
    else {
        print "$_ in both files\n";
        }
    }
[download]

Since I usually get passed excel files, I haven't had the occassion to work with 190,000 line files. I haven't had any slowness issues, though.