There is a lot of good advice in the previous posts and I think that you should take it into account.

However, apart from gellyfish, they also seem to have missed the point. ;-) To avoid duplicates in "NEWFILE" you could do something like the following:

#!/usr/local/bin/perl -w

use strict;

open REGFILE,     "projects.dat" or die "Error message here: $!\n";
open NEWFILE, "+>>subscribe.dat" or die "Error message here: $!\n";

# Rewind the file for reading
seek(NEWFILE, 0, 0);

# Store all lines in NEWFILE in %seen using a hash slice
my %seen;
@seen{<NEWFILE>} = undef;

while (<REGFILE>) {
    if (not exists $seen{$_}) {
        $seen{$_} = undef;
        # Do your split etc here
        print NEWFILE $_;
    }
}

close (NEWFILE);
close (REGFILE);
[download]

--
John.