1) Presuming you'll be running this more than once, using a hash in the sense you're implying may not help. Since existing albums directories will be present, you need to check for their existence anyway before trying to create them. That can be done with the filetest -d, which implicitly checks for existence as well directory-ness. Just test the directory's existence before attempting to create it each time.

Using a hash for the files that you *are* moving might speed up your program if you have many tracks going into the same directory, but you'd be lucky to see an improvement, even in a benchmark. Besides, speed doesn't appear to be your primary concern here.

Also, where are you getting the album information? Your snippet doesn't get it, though MP3::Info allows you to retrieve it from the same hashref.

2) This sounds more like a design problem than a Perl problem. You need to decide whether you want to sort tracks based on album or artist. Personally, I sort my MP3s by artist then album (i.e. two dirs deep). And yes, soundtracks are difficult. I just have a Soundtrack directory and a directory under that for each movie title.

Perhaps this (untested) code will help get you on track.

#!/usr/bin/perl -w
use strict;

use MP3::Info;
use File::Copy;
use File::Find;
use File::Basename;

find(\&wanted, "/MP3");

sub wanted
{
    /\.mp3$/ or return;
    my $tag = &MP3::Info::get_mp3tag($File::Find::name,1,1);
    my $artist = $tag->{ARTIST};
    my $album = $tag->{ALBUM};

    $artist |= "Unknown Artist";
    $album |= "Unknown Album";

    $artist =~ s/ /\\ /g;
    $album =~ s/ /\\ /g;
    
    my $dir = $File::Find::dir;

    -d "$artist" or mkdir "$dir/$artist" 
      or die "Couldn't create $dir/$artist: $!";
    -d "$artist/$album" or mkdir "$dir/$artist/$album"
      or die "Couldn't create $dir/$artist/$album: $!";

    move $File::Find::name, "$dir/$artist/$album";
}
[download]