Advent of Code 2019 - Day 3

JEG2 · December 3, 2019, 10:55pm

sasajuric · December 3, 2019, 11:14pm

I like the intensive usage of streams

alvises · December 3, 2019, 11:41pm

qhwa · December 4, 2019, 12:54am

Interesting solution. I like the idea of putting all points in two MapSets, the code is pretty straightforward to read.

patrickcarver · December 4, 2019, 2:48am

My solution

JEG2 · December 4, 2019, 3:40am

Funny you say that, I did actually say to myself when I sat down that I wanted to try your technique of stream-all-the-things today. In the end though, I didn’t feel like I fit it to the problem very well.

I rarely bother to explain my solutions because I post them so late in the cycle that my ideas are already represented. I do read through the solutions of others when I’m done though.

albertarv · December 4, 2019, 6:29am

I ended up switching datastructure two times for day 3. Started with lists for the wires, but switched to MapSet to be able to use MapSet.intersection/2. Then for part 2 I switched to Map to be able to track steps, but continued comparing using MapSet.intersection by first converting to a MapSet.

My solution is a bit overkill with nested modules to contain different parts of the code, it’s something I’ve experimented with recently, to see if the code stays more readable when I return to it later. Though I didn’t go all the way through with it. Code left it as it was when I finished part 2.

github.com

scmx/advent-of-code-2019-elixir/blob/master/lib/day_03_crossed_wires.ex

defmodule Adventofcode.Day03CrossedWires do
  use Adventofcode

  alias __MODULE__.{CentralPort, Parser, Twister, Wire}

  def part_1(input) do
    closest_intersection_distance(input)
  end

  def part_2(input) do
    fewest_combined_steps_to_intersection(input)
  end

  def closest_intersection_distance(input) do
    input
    |> Parser.parse()
    |> Twister.twist()
    |> intersections()
    |> Map.keys()
    |> Enum.map(&manhattan_distance({0, 0}, &1))

This file has been truncated. show original

sasajuric · December 4, 2019, 9:23am

Yeah, it doesn’t seem that streams are optimal here. To be clear, I don’t use streams that much in real code, but I like to use these challenges to practice them a bit.

michallepicki · December 8, 2019, 1:32pm

Escripts are slow and my initial solution with maps was very slow so I re-wrote my Gleam solution after watching this: https://github.com/michallepicki/adventofcode/blob/master/2019/gleam/src/crossed_wires_after_watching_freds_recording.gleam

(could optimize even further but meh)

I’m slowly catching up!

simon · December 8, 2019, 7:59pm

Finally solved Day03 after tracking down a silly logic error which didn’t show up in any of the tests but did when I was trying to calculate my part 2 result. Code is here.