So, yesterday I was chatting with my daughter. She was talking with her team and for some reason one of them pulled out an interview question from their company's question bank. Turns out it was today's Advent of Code problem.

As with past years, I'm going to try to solve the problems in Clojure but if I can will talk Python when I talk about solutions.

Part 1 of the problem basically asks for you to find a pair of numbers in an array that sum to a specific value. In this case 2020. Figuring that this was day 1 I didn't expect a crazy large data set or any other tricks or weirdness - a straightforward solution should suffice. Simply a loop within a loop and sum up all the pairs:

1
2
3
4

for x in data:
  for y in data:
    if x+y == 2020:
      print(x*y)

Nothing fancy but it gets the job done. It does print out the answer twice since x and y both go through all the indices but that's no big deal.

The part 2 twist was that now you were looking for a set of three entries that added up to 220. Once again, not a big deal:

1
2
3
4
5

for x in data:
    for y in data:
        for z in data:
            if x+y+z == 2020:
                print(x*y*z)

As before, you'll get multiple answers but no big deal.

What I like about this one is that you can think about this as a looping exercise as above but you can also think about it as a list processing exercise, that is, by thinking about it with more of a functional programming bent.

The key insight here is that the question was clear in that there will only be one pair in part 1 that satisfies the problem and likewise only one pair in part 2.

If we look at each item in our data set, it's part of the answer if and only if there's another number in the set equal to 2020 minus that item. This leads to a list comprehension

1

part1_list = [x for x in data if 2020 - x in data]

Now, part1_list should contain the two items we need. The first value x was found when the comprehension saw that 2020 - x was in the list and 2020-x which is the second value was confirmed when the for part of the comprehension gets to it and finds that the first item is in the list. Then, it's a simple matter of just multiplying the two numbers together for the answer.

Part 2 is similar but you can use a list comprehension to iterate over all pairs of elements and then you calculate the third:

1

part2_list = [x for x in data for y in data if 2020-(x+y) in data]

Of course you could have put y or 2020-(x+y) in place of that first x.

Part 2 has an additional subtlety in that you'll get the solution multiple times which makes sense you're hitting each triple multiple times. To fix that, turn it into a set:

1

part2_set = set(part2_list)

and then calculate the product.

At the core, both of these solutions are really the same but you get to them by thinking very differently. The first one is all about the loops - thinking about data[i] at a very low discrete level. The second approach is thinking about the data as a list and processing that list at a much higher level. This could be an ice problem to transition between the two approaches.

Looking forward to what tomorrow's problem brings.