To be honest with you, I would be reaching for some hashes at this point. Given the relationship you have described, I would create a hash that looks kinda like this:

# Sorry about the MANY keyword - I would choose a better name
# for the secondary data, but I have no clue what would make
# sense.
my %keyhash = ( $id => { NAME => $name,
                 REF  => $ref,
                         MANY => []
                       }
              );
[download]

To populate it, I would do something like this:

for ( @array1 ) {
  my ($id,$name,$ref) = split /,/;
  # You may wish to add some error checking to make sure the
  # hash key $id does not already exist
  $keyhash{$id} = { NAME => $name,
            REF  => $ref,
            MANY => [],
                  }
}

for ( @array2 ) {
  my ($id,$name,$ref) = split /,/;
  # Warn and do nothing if a record is found for which the
  # $id is not already in %keyhash
  unless ( defined( $keyhash{$id} ) ) {
    warn "No such record $id!\n";
    next;
  }
  push @{$keyhash{$id}{MANY}}, [ $name, $ref ];
}
[download]

Although you may want MANY to be a hash - it really depends on how you want to use your data later.

Finally, to extract the number of records for each ID,

# A little something to get the plurality correct
for ( keys %keyhash ) {
  my $num = @{$keyhash{$_}{MANY}};
  printf "%s appeared %d %s\n",
    $_,
    $num,
    $num > 1 ? "times" : "time";
}
[download]

I will say my choice of data structures ( hashes instead of arrays ) is really dependant on how you intend using the data later. Hashes, for me, seem to better reflect the relationship between tables better than arrays. YMMV, of course.

Updated 14:04 It was pointed out I had dropped an equals sign. Sigh.

mikfire