No comments on Heap not having ever used it.
However, unless you are dynamically adding and removing items to your dataset wouldn't something like:
sub smallest_n (&\@$) { my ($cmp, $arrayref, $n) = @_; return unless $n && @$arrayref; $n = @$arrayref if @$arrayref < $n; my @results = sort $cmp @$arrayref[0..$n-1]; local ($a, $b); $a = pop @results; for (my $i = $n; $i < @$arrayref; $i++) { $b = $arrayref->[$i]; if ($cmp->() == 1) { @results = sort $cmp (@results, $b); $a = pop @results; }; }; return (@results, $a); }; use Test::More tests => 1; use List::Util qw(shuffle); my @a = shuffle (1..100000); my @b = smallest_n {$a <=> $b} @a, 5; is_deeply [@b], [1,2,3,4,5];
be easier? The support for the heap datastructure will add a fair bit of overhead with your large dataset, so I wouldn't use one unless you need it.
Update: totally misready what blakem was proposing. D'oh.
In reply to Re: Heap sorting in perl
by adrianh
in thread Heap sorting in perl
by blakem
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |