cut -f3 -d, <file1.csv | sort -u >file1.cut.sort
cut -f3 -d, <file2.csv | sort -u >file2.cut.sort
comm -23 file1.cut.sort file2.cut.sort
[download]

use strict;
use warnings;
my $file2 = {};
open my $fh2, "<file2.csv";
while( <$fh2> ) {
    chomp;
    my @fld = split /\,/;
    $file2 -> { $fld[2] } = 1;
}
close $fh2;
open my $fh1 "<file1.csv";
while( <$fh1> _ ) {
   chomp;
   my @fld = split /\,/;
   defined( $file2 -> { $fld[2] } ) and next;
   print "$_\n";
}
close $fh1;
[download]

Handy code, but it does suffer the problem that the commas might be quoted or that there are newlines in quotes (a common problem with hand-rolled CSV parsers). If the OPs CSV files are simple enough, then your code will be useful. Otherwise, he or she may be forced to find a way to sneak in modules (which is often much easier than it sounds).

In any event, take my comments as caveats, not criticisms :) Your code has a good chance of working for simple CSV files.

Cheers,
Ovid

New address of my CGI Course.

he or she may be forced to find a way to sneak in modules

Or, alternatively, try the standard Text::ParseWords module. I've never understood why that module isn't used more widely.

--
<http://dave.org.uk>

"The first rule of Perl club is you do not talk about Perl club."
-- Chip Salzenberg

Very true. I think I once wrote a simple Split routine that walks through strings willy-nilly where I was in a "no module" situation like the OP combined with applicability of the caveat you describe.

-M

Free your mind