Hello Speed_Freak, you question is confusing me: too much data, no code at all, no code from your part, no expected results and I do not really well understand this subgroups and the goal..

But since you are asking where to start.. know your data is a good suggestion and and another good quote sounds like: when you know deeply your data, then algorithm is a matter of simply implementation.

So where to start? ordering => array and indexing => hash

I mean that when you are processing your data you split up elements and fill a datastructure that suits your needs. So the basic is a simple loop that consumes lines of data:

use strict;
use warnings;

while (<DATA>){
  chomp;
  my @ele = split /\s/,$_;
[download]

Now that you has @ele you need to coherce it to your logic: so supposing you need to store which ID ( $ele[0] ) has $ele[1] + $ele[2] you can indexing the $ele[1] $ele[2] presence and use it as key of an hash and pushing IDs as values of an anonymous array:

use strict;
use warnings;

my %res;
while (<DATA>){
  chomp;
  my @ele = split /\s/,$_;

  push @{ $res{"$ele[1] $ele[2]"} }, $ele[0];
}
__DATA__
1 monkey cow hammer nail
2 monkey sheep hammer nail
3 dog cat hammer nail
4 monkey cow hammer nail
[download]

this leads you to a datastructure like: ("dog cat", [3], "monkey sheep", [2], "monkey cow", [1, 4])

If you just need to know which ID has monkey you'll loop keys of the hash searching the pattern monkey as in:

foreach my $key (keys %res){

  if ($key =~ /monkey/) {

     print "monkey [occurence in $key] found in IDs:", (join ', ', @{$
+res{$key}}), "\n";
[download]

This is my where to start

L*

PS perldsc and (2004)Using Perl for Statistics: Data Processing and Statistical Computing as readmore suggestions.

L*