Hi nirelit,

a more detailed description on your approach with the reverse index would be also much appreciated

Please mark updates to your nodes as such. See How do I change/delete my post?, especially "It is uncool to update a node in a way that renders replies confusing or meaningless".

Anyway, I'll give you the following hint: as you're reading your input files, let's say you have the current directory name stored in the variable $curdir, and you're reading your input file line-by-line (see for example "Files and I/O" in perlintro). Then you could do something like this:

my %reversetbl;  # at the beginning

# ... open each file ...
while (<$fh>) {
    chomp; # remove newline
    # ...
    $reversetbl{$_}{$curdir}++;
}

# at the end:
for my $k (keys %reversetbl) {
    print $k, ';', join(',', keys %{ $reversetbl{$k} } ), "\n";
}
[download]

This will give you the "6576576;dir1,dir2" output you want.

To see what this code is doing, you can look at the data structure (a "hash of hashes") with Data::Dumper, for example: use Data::Dumper; print Dumper(\%reversetbl);.

Hope this helps,
-- Hauke D