my $p = HTML::TokeParser::Simple->new(\$html); my (%href, $this_href, $number, $letter); while (my $t = $p->get_token){ if ($t->is_start_tag('h2')){ $letter = $p->get_trimmed_text('/h2'); next; } if ($t->is_start_tag('a')){ # skip bookmarks next if $t->get_attr('name'); $this_href = $t->get_attr('href'); next; } if ($t->is_start_tag('span')){ $number = $p->get_trimmed_text('/span'); $href{$letter}{$this_href} = $number; next; } }