Re: Two dimensional sets intersection

[Not_a_Number]

No need to worry about narrowing your search or whatever. Hash lookups are fast.

The code below creates a hash for a 100 x 100 matrix. It also assumes that each matrix pair has a value (worst case, given your 'spec'). To simplify validation, I make this value 1, but changing it to any other number makes no difference to the time required for lookup.

Subsequently, I make another worst-case assumption, namely that your lists of search terms (what you call Subset 1 and Subset 2) are maximally large.

On my machine, it finishes before I can blink. Run it and see:

use strict;
use warnings;

# Create 100 x 100 datastructure:

my @uc = ( 'AA' .. 'DV' );
my @lc = ( 'aa' .. 'dv' );

my %index;
for my $k ( @uc ) {
  for my $v ( @lc ) {
    $index{$k}{$v} = 1;
  }
}

# Parse it:

my $score;
for my $k ( @uc ) {    # or: for my $k ( @subset_1 )
  for my $v ( @lc ) {  # or: for my $v ( @subset_2 )
    $score += $index{$k}{$v} || 0;
  }
}

print $score;