A couple of (uncalled for ;--) comments on your script:

• if a system function (open) fails you have to die, just using || "error" will not help,
• $! is the error description, it is usally a good idea to display it when a system function fails,
• to test existence is usally done using a hash.

A couple of style points:

• if you want to open a file for reading (regfile), why not say so in the open: open( FH, "<file") instead of using >>, BTW usually filehandles are in CAPS,
• you can use undef as a placeholder in split,
• there is no need to escape . and | in strings,
• instead of reading a whole file in an array and then processing each line why not read each line, process it and then go to the next one? This way you don't have to slurp the whole file in memory.

So your code modified:

#!/bin/perl -w
use strict;

open (REGFILE, "<$datadir/projects.dat") || die "Cannot open File1: $!
+";

open (NEWFILE,">>$datadir/subscribe\.dat") || die "Cannot open File2: 
+$!";

my %seen_email; # hash email => 1
while( <REGFILE>) {
        my( $email, $name, $format);
        ($email, $name, $format, undef) = split /\|/;
        next if( $seen_email{$email});
        print NEWFILE"$email|$name|$format\n";
        $seen_email{$email}=1; # either just the email or the whole li
+ne
}

close (NEWFILE);
close REGFILE;
[download]

An other option, if you are on any kind of unix (or cygwin) is to preprocess the initial file using sort -u:

#!/bin/perl -w
use strict;

# the die is actually useless here
open (REGFILE, "sort -u $datadir/projects.dat |") || die "Cannot open 
+File1: $!";

open (NEWFILE,">>$datadir/subscribe\.dat") || die "Cannot open File2: 
+$!";

while( <REGFILE>) {
        my( $email, $name, $format);
        ($email, $name, $format, undef) = split /\|/;
        print NEWFILE"$email|$name|$format\n";
}

close (NEWFILE);
close( REGFILE) || die "could not sort $datadir/projects.dat: $!"; # i
+n this case the error happens when you close the file
[download]