Advent of Code 2023 Day 7 (BaseA Style)

AlteryxCommunityTeam

Discussion thread for day 7 of the Advent of Code - https://adventofcode.com/2023/day/7

Advent of Code

Accepted answers

All comments

jarrodthuener

Not the prettiest, but it works. Had fun thinking through Part 2 combinations

Spoiler

Day7.yxzp

Gawa

There were a pit fall for me even in part 1, and took a bit long but get it done

Spoiler

AoC2023_Day7.yxmd

CoG

Tried to pretty up my solutions before posting (still not lookers, but much better than when they started):

Spoiler

Solution.yxmd

ScottLewis

I usually try to bash through these without planning, which is terrible form in the real world but this isn't there. I end up with ugly but functional workflows with a lot more pieces than would be needed if I took my time. Lots of path for each case where a well applied multifield/tile/etc. could handle many cases at once. Still, work is work:

Day7Part2ScottLewis.PNG

AncientPandaman

keep forget the requirement...

Spoiler

forget sort by A > K > Q > .. > 2
forget sort descending
forget part 2: is A > ... 2 > J and etc...

workflow
Screenshot 2023-12-07 142933.png

tokenize and group by cards, then crosstab with count of card.
it show as below, each column is count of 1 of kind, 2 of kind, 3 of kind, 4 of kind and 5 of kind. (joker is part 2, split for ease calculate)

e.g. count_2 = 1 mean 1 pairs or 2 cards.
e.g. count_3 = 1 mean 1 three of kind or 3 cards.
Screenshot 2023-12-07 143401.png

for sorting, i create a table then use find and replace to replace with card value. for ease of sort later.
Screenshot 2023-12-07 143020.png

example result before sum, the "sorting" column is converted from table above for sorting.
Screenshot 2023-12-07 143212.png

for part 2, I try to generate a formula for it. but i think it better to manual generate a table for each kind of situation. and it work perfectly.
Screenshot 2023-12-07 143008.png

Day 07.yxmd

DaisukeTsuchiya

Today's one is interesting.

Spoiler

Watermark

The 2 are almost the same, slight tweak for an extra stream to handle the Jacks.

I handled it essentially as 2 pieces. 'hand ranking', ignoring what the cards 'are' only how many of the same. Then 'sorting' based on actual card values.

Spoiler

AOC 2023 Day 7P2.yxmd

Tokimatsu

I misunderstood the requirement.

Spoiler

I thought that with pairs, the one with the stronger pair number would win.
スクリーンショット 2023-12-07 160600.png

cgoodman

Spoiler

I struggled with the question logic, as I was trying to work out why if K>T why they would be ordered the way they did, so lots of re-reading and working it out.

My approach was:
1) Split the cards to rows then assign a value 1-12 for 2 - A
2) Map the cards to the type of hand for example 5 card of 1 type = five of a kind, 3 of one card and 2 individuals were three of a kind etc, to give the hand strength
3) Replace TJQKA with ABCDE with the original hand order which when combined with hand strength gave the ability to sort matching hands based on the strongest first card (again was struggling with the logic compared to real poker)

For part 2
1) Isolated hands with Jacks, then appended all combinations and followed the same logic as above to find the strongest hand
2) Replaced this with the original set
3) Modified Jacks to 0 instead of B for the sorting

kelsey_kincaid

This was a really fun puzzle!

Spoiler

Part 1: I overcomplicated this at first, until I realized I could simply sort a cross tab to get the ranks of each hand.

Part 2: Mostly a copy and paste from part 1. I could go back and clean this up so that it's a parameterized workflow instead of two separate workflows for each part.... but AOC isn't always about elegance

Luckily was able to determine early that forming the best hand just means adding the number of jokers to the card with the highest frequency & strength in each hand. My lone JJJJJ hand dropped off in a join, which threw off my answer until I looked more critically at my record count.

AOC2023_Day7_KelseyKincaid.yxzp

alisonpitt

I've pretty much given up on trying use fewer tools, at this point 😅 But this was a really fun one, I'm really pleased I got it!

Spoiler

Edit: it's fascinating to see how many different approaches there are in this thread!!

Day 7 - AP.yxmd

JeffF

Spoiler

AkimasaKajitani

My solution.

Today, I couldn't perform as well as I could because I have some work. So, I took so much time for part 2 due to a small mistake.

Spoiler

Day07.yxzp

Yoshiro_Fujimori

My solution.

Spoiler

AoC07.yxmd

AkimasaKajitani

My Designer Cloud solution.

Spoiler

I was surprised at the difference of the Find Replace tool.

And I optimize my Designer Desktop workflow.

Spoiler

I found that there were some waste tools after upload, I changed the screenshot.

Day07_2.yxzp

DataNath

Not winning any tool golf here but very much enjoyed today's challenge! Part 2 was interesting to work through.

Spoiler

Day 7.yxmd

cjaneczko

I thought I had a good workflow going until I found buried in the documentation that Mode calculation selects the lowest number when there is a tie. This is a good use case where Alteryx should have an option to pick MaxMode or MinMode in the Summarize Tool when there are more than one Mode.

patrick_digan

Probably not the straightest path there but it worked.

Spoiler

Day7.yxmd

Kenda

Spoiler

7.yxmd

7 pt 2.yxmd

Kenda

@patrick_digan wrote:
Probably not the straightest path there but it worked.

If you use Ctrl Shift + it would be a straighter path 😉

rachel_k

Spoiler

It's not pretty, but it worked. I edited part 2 right on top of part 1, so this won't work for part 1.

day7.yxmd

FrederikE

Spoiler

07.yxmd

mmontgomery

Tried to get too cute on p2 and simplified it.

Spoiler

2023_Day7.yxzp

StevenP

My Solution:

Spoiler

I applied a ranking to the hand types and each card value and did nested sorts within each to determine the true order.

In order to account for Jokers I looked at the relevant hands and determined the best card to "add" the joker into.
For example: the hand KTJJT has three different cards K/J/T with T being the most frequent card with 2 instances. Taking the max value of 2 and adding the count of jokers (2) to it results in a count of 4 making it a four of a kind. To account for the fact that a native four of a kind is always better than one that uses jokers I did not convert the original hand to be KTTTT but left it as is. The Type_Code however was updated from Two Pair (2,2,1) to Four of a Kind (4,1).

Day 7 Camel Cards.yxmd

ccanterman

Solved! That was a fun one. I ended up with a very bizarre detour with a couple of Multi-Rows to handle the Jokers and getting them sorted properly

Spoiler

phottovy

Fortunately I was able to use most of my solution for part 1 for part 2. Detour tools for the win!

Spoiler

aoc_ph_2023_07.yxzp

mceleavey

Another easy one today...

Spoiler

It's not pretty, but I went down the route of splitting out the data into it's parts, then tokenizing to rows on the cards by individual characters, so I had a card per row. I then grouped by Hand and by card to get a count of each card per hand. I then filtered (many times) to determine if they had 5,4,3 or 2 of any kind in a hand (this process had to be repeated).
This got me to a point where I determined the hand rank (1 being 5 of a kind and so on).
I then brought in the card scores from a text input to join. From here it was a case of ranking them in Hand ranking order, then sorting by 1,2,3,4 and 5 in order by card score.
I then used two record IDs with an order shift in between as a quick way of scoring highest to lowest (could have done that earlier)

For part two I used a formula to determine the hand shift according to no. of Jacks:

if [No. Jacks]=0 THEN [Hand Rank]

ELSEIF [Hand Rank]=1 then 1

elseif [Hand Rank]=2 AND [No. Jacks]>0 then 1

elseif [Hand Rank]=3 AND [No. Jacks]>0 then 1

elseif [Hand Rank]=4 AND [No. Jacks]>0 THEN 2

elseif [Hand Rank]=5 AND [No. Jacks]=1 then 3
elseif [Hand Rank]=5 AND [No. Jacks]=2 then 2

elseif [Hand Rank]=6 AND [No. Jacks]=1 then 4
elseif [Hand Rank]=6 AND [No. Jacks]=2 then 4

elseif [Hand Rank]=7 AND [No. Jacks]=1 then 6

ELSE [Hand Rank] ENDIF

and the card score was adjusted in the text input.

So there you go, @phottovy , enjoy!

Samantha_Jayne

Spoiler

ok I am busy closing deals and rushed this one. #sorrynotsorry
Don't try to understand it, ill clean it up another day and upload!
Good news is though I got the right answer in all that spaghetti 🤣