Project Euler #17, Take Three

So, this version starts with the last version and applies a bit of smarts. Essentially, I said to myself, "Hey! The number of letters in the numbers 100 through 199 are the same as the number of letters in 1 through 99 plus 100 times the number of letters in "one hundred". I reworked the code to take advantage of that. Then I realized that the NumberLength function only needed to work on numbers between 1 and 99 inclusive, and reworked that as a special case. And as soon as I did that, it was obvious that NumberLength only had two cases, one for numbers between 1 and 20, one for 21 to 99, and just made two loops each of which had the appropriate piece of NumberLength code inside.

	use v6;
	my %number_lengths = (
	0 => 0, # special case -- we only hit this one as the $ones on a multiple of ten
	1 => 3,
	2 => 3,
	3 => 5,
	4 => 4,
	5 => 4,
	6 => 3,
	7 => 5,
	8 => 5,
	9 => 4,
	10 => 3,
	11 => 6,
	12 => 6,
	13 => 8,
	14 => 8,
	15 => 7,
	16 => 7,
	17 => 9,
	18 => 8,
	19 => 8,
	20 => 6,
	30 => 6,
	40 => 5,
	50 => 5,
	60 => 5,
	70 => 7,
	80 => 6,
	90 => 6
	);
	my $one_to_ninety_nine = 0;
	for 1..19 -> $number
	{
	$one_to_ninety_nine += %number_lengths{$number};
	}
	for 20..99 -> $number
	{
	my $ones = $number % 10;
	$one_to_ninety_nine += %number_lengths{$number - $ones} + %number_lengths{$ones};
	}
	my $letter = 10 * $one_to_ninety_nine;
	for 1..9 -> $hundreds
	{
	# 7 below is for "hundred"
	$letter += (%number_lengths{$hundreds} + 7) * 100;
	}
	$letter += 11; # 11 is for "one thousand"
	say $letter;

view raw euler_017_03.pl hosted with ❤ by GitHub

The resulting code runs in just a shade under two seconds on my MBP. That's a solid 90x improvement over my initial code. But Rakudo's overhead for starting the script is over one second. If you subtract that from all runs, this current version is under a second, and very close to be 200x faster than the original script.