Lots of time, a good data structure is the key. In this case, use a HOA to group nodes.

use Data::Dumper;
use strict;
use warnings;

open DATA, "<", "foo.dat";
my $data = {};
while (my $line = <DATA>) {
    my @columns = split /,/, $line;
    push @{$data->{$columns[0]}}, [$columns[1], $columns[2]]; 
}
close DATA;
print Dumper($data);
[download]

With the sample data you give, here is the output:

$VAR1 = {
          'AH2S21003' => [
                           [
                             '2004-01-16 02:23:05.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-16 02:24:05.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-16 02:24:05.000000',
                             'ANE4987E Error processing
'
                           ]
                         ],
          'AH2D21001' => [
                           [
                             '2004-01-15 22:57:32.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-15 22:57:33.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-15 22:57:34.000000',
                             'ANE4987E Error processing
'
                           ]
                         ],
          'ESI2A55P' => [
                          [
                            '2004-01-16 04:21:43.000000',
                            'ANE4037E File Skipped
'
                          ],
                          [
                            '2004-01-16 04:25:43.000000',
                            'ANE4037E File Skipped
'
                          ],
                          [
                            '2004-01-16 04:27:43.000000',
                            'ANE4037E File Skipped'
                          ]
                        ],
          'ABHS00001' => [
                           [
                             '2004-01-16 01:43:24.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-16 01:46:24.000000',
                             'ANE4987E Error processing
'
                           ],
                           [
                             '2004-01-16 01:49:24.000000',
                             'ANE4987E Error processing
'
                           ]
                         ]
        };
[download]

That's a good way to store the data structure (I was going to reply and suggest a hashref of arrayrefs). I'd just like to add a caveat: be careful, slurping an entire file into memory can be a bad idea if the file is too big -- especially if you're going to manipulate the data.

A better bet may be to create an array of filehandles, append the data to each filehandle, and the reread in to reduce the total amount of memory you're going to need.

-Dan

Um ... I think this is what you're asking for ...

while (<>) {
   my ($node) = split /,/;
   open $node, ">>$node" or die "a horrible death!";
   print $node $_;
   close $node;
}
[download]

There are slicker ways though. Other than opening a file each time, you could cache the opened filehandles and just write to the already opened ones (mayn't be too useful if you've got many many files to open) Here's some code (untested, caveat lector):

my %cache;
while (<>) {
   my ($node) = split /,/;
   unless ($cache{$node}) {
      open $node,">>$node" or die "a horrible death";
      $cache{$node}++;
   }
   print $node $_;
}
for my $n (keys %cache) { close $n; }
[download]

#!/usr/bin/perl

use strict;
use warnings;

my $file = "";
my $fh;

while (<>) {
    my ($node, $mess) = split /,/ => $_, 2;
    unless ($node eq $file) {
        close $fh or die "close $file: $!" if $fh;
        $file = $node;
        open $fh => '>', $file or die "open $file: $!";
    }
    print $fh $mess;
}

__END__
[download]

Or from the command line:

perl -apF, -e'if($F ne$F[0]){$F=$F[0];open F,">$F"or die;select F}s/[^
+,]+//'
[download]

Abigail

I have to agree.

dbrock, is there a reason you're dumping the select to a file first?

while(<>) {
    my @a = split /,/;
    $files{pop @a} = @a;
}
[download]