OK, I got a working version. I things around as you suggested (hopefully). Now I want to make sure it is as efficient as possible. Can you please look at this code and see if it can get anymore efficient. or should I open a new thread?

sub deDuppedArray
{
    my ($a_ref, $b_ref, $my_bonly, $my_isec_plus_bonly) = @_; # refere
+nce
    my (%Aseen, %Bseen, %count, @isec, @diff, @union, @aonly) = ();
    @Aseen{@$a_ref} = (); # lookup table
    @Bseen{@$b_ref} = (); # lookup table
    foreach my $e (@$a_ref, @$b_ref) { $count{$e}++ } # put all items 
+in hash table
    foreach my $e (keys %count)
    { # interate over each key of hash table
        push(@union, $e); # keys of hash table = union
        if ($count{$e} == 2)
        {
            push @isec, $e; # seen more than once = intersection
        } else
        {
            push(@aonly, $e) unless exists $Bseen{$e}; # seen once + f
+rom A = Aonly
            push(@$my_bonly, $e) unless exists $Aseen{$e}; # seen once
+ + from A = Aonly
        }
    }
    @$my_isec_plus_bonly = ( @isec, @$my_bonly);
}
[download]