I had a similar task the other day (although it was dbm, it could just as well have been a text file). I had a bunch of information in these files that I needed to strip out and place in a hash -- namely county codes and their corresponding email aliases. Sounds like you would just need to modify this by using opendir / readdir etc as pointed out by chanio and prasadbabu to avoid hard coding the file name(s) and allow traversing the directory.

Post some of your code and we can help you more, til then here is a little snippet that can give you something to start with:

#!/usr/bin/perl -wT
dbmopen(%ISO_Country, "ISO_Country", 0777)
    or die "Cannot create ISO_Country: $1\n\n";

dbmopen(%ISO_Email, "ISO_Email", 0777)
    or die "Cannot create ISO_Email: $1\n\n";

open FNAME, $ARGV[0]
    or die "Cannot locate file $ARGV[0]: $1";

while ($line = <FNAME>){
    chop($line);
    @data = split /\t/, $line;
    $ISO_Country{$data[0]} = $data[1];
    $ISO_Email{$data[0]}   = $data[2];
}

&displayData;

dbmclose(%ISO_Country);
dbmclose(%ISO_Email);

sub displayData{
    while (($key, $value) = each %ISO_Country){
        print "ISO($key), Country($value)\n";
    }
    while (($key, $value) = each %ISO_Email){
        print "ISO($key), Email($value)\n";
    }
}
...
[download]