in reply to Sampling from Combination Space
You can get a random subset without replacement using List::Util::shuffle().
To build a sample of picks with repetition allowed:use List::Util qw/shuffle/; sub pick { return if @_ != 2; my ($num, $count) = @_; (shuffle 1 .. $num)[0 .. $count-1]; }
You can keep them unique by storing the samples as keys of a hash,my ($top, $ct, @samples) = (158, 8); my $size = 50,000; $#samples = $size; $_ = [pick( $top, $ct)] for @samples;
If you need a memory-breaking large sample, look to a file or database to hold it. Tie::File would be simple to use for that.my %sample; $sample{join ',', pick($top, $ct)} = undef while $size > keys %sample;
Does order matter in a sample? If not, then sort them to get a canonical order before storage.
After Compline,
Zaxo
|
|---|