Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:
Howdy all,
I have a function that I use sometimes when I need permutations (I do permutation-testing often for stats problems). Currently I require the user to pass the "cardinality": how many different elements to put in each class. Let me give a simple example first, before I show the code:
Data: a1, b2, c3 Cardinality: 1 Permutations: {a1, b2}, {a1, c3}, {b2, c3}, Data: a1, b2, c3, d4 Cardinality: 2 Permutations: {{a1, b2}, {c3, d4}} {{a1, c3}, {b2, d4}} {{a1, d4}, {c3, d4}}
I hope that's reasonably clear: cardinality defines the size of the group I wish to obtain permutations of. The first example above was: all possible pairs of groups of one element. The second example was: all possible pairs of groups of three elements (triplets). Within a single permutation, no element can be repeated, and I take only unique permutations.
I am generating all these permutations into a data-structure for later use (e.g. passing them to other software, writing them to file with their results, sending them to a DB, etc.). The code that does this is pretty simple, but unfortunately it has required me to HARDCODE the cardinality. It looks like this:
my @data; if ($cardinality == 1) { for (my $i = 0; $i < scalar(@$permlist); $i++) { for (my $j = 0; $j < scalar(@$permlist); $j++) { push @data, { first => [$$permlist[$i]], second => [$$permlist[$j]], }; } } } if ($cardinality == 2) { for (my $i = 0; $i < scalar(@$permlist); $i++) { for (my $j = 0; $j < scalar(@$permlist); $j++) { for (my $k = 0; $k < scalar(@$permlist); $k++) { for (my $p = 0; $p < scalar(@$permlist); $p++) { push @data, { first => [$$permlist[$i], $$permlist[$j]], second => [$$permlist[$k], $$permlist[$p]], }; } } } } }
And so on for higher cardinality. I would prefer to have a generalized way of doing this, rather than needing a separate set of for-loops for each cardinality. Is that possible? I'm turned it over in my head, and haven't found a way yet.
|
---|
Replies are listed 'Best First'. | |
---|---|
Re: Generalizing Code: Generating Unique Permutations
by diotalevi (Canon) on Oct 07, 2003 at 19:57 UTC | |
Re: Generalizing Code: Generating Unique Permutations (iterator)
by tye (Sage) on Oct 07, 2003 at 21:42 UTC | |
Re: Generalizing Code: Generating Unique Permutations
by Anonymous Monk on Oct 07, 2003 at 21:58 UTC | |
Re: Generalizing Code: Generating Unique Permutations
by dragonchild (Archbishop) on Oct 07, 2003 at 20:16 UTC | |
Re: Generalizing Code: Generating Unique Permutations
by ig (Vicar) on Dec 17, 2008 at 05:34 UTC |