I've made a couple of assumptions in my answer.

1. Your data has no embedded commas (if it has you'll need to look at Text::CSV_XS).
2. You are using columns 0, 1 and 2 from the data. You can adjust this in the line with the call to split.

Here's the code:

open(DAT, 'file.dat') or die "Can't open file.dat: $!\n";

my %data;
while (<DAT>) {
    chomp;
    my ($key, @coords) = (split(/,/))[2, 0, 1];

    push @{$data{$key}}, \@coords;
}

foreach (keys %data) {
    print "$_\n";

    local $" = ', ';
    foreach my $coord (@{$data{$_}}) {
    print "@$coord\n";
    }
    print "end\n";
}
print "end\n";
[download]

Basically I'm building up a hash where the key is grouping column and the value is an array of co-ordinates. Having read the whole file and built up the hash, I then iterate over the hash again and print out hte values.

--
<http://www.dave.org.uk>

European Perl Conference - Sept 22/24 2000
<http://www.yapc.org/Europe/>