Advent of Code 2024 Day 9 (BaseA Style)

AlteryxCommunityTeam

Discussion thread for day 9 of the Advent of Code - https://adventofcode.com/2024/day/9

Advent of Code

Accepted answers

All comments

AncientPandaman

first?

Spoiler

part1:
split to the step accordingly then create a freeID in ascend and fileID in descend. verified if the positionID will be smaller.

part2:
use data given before split to the line. (keep the data size, to easier to compare later)
check one by one from bottom, if size is enough, position is earlier.
generate another rows if free size is larger.

loop for 10k iteration, and take 40 mins to run it
Screenshot 2024-12-09 143011.png

Day 09.yxzp

CoG

There are many optimizations that can be made here to shorten time needed for Part 2 (I have only used a few, but have a V2 that I'm working on). Current workflow runs in ~20 minutes:

Spoiler

A lot of joins and data manipulation, but fairly straight forward. The optimization I used here (for part 2) is to exclude records that are on the right and no longer being considered or so far to the left that no free space exists to move anything else to. Keeping these records does nothing for the calculation but expend memory.

Iterative Macro (Part 2 - File Manager):
File Manager.png

V2 (Part 2 runs in ~8 minutes):

Spoiler

Primary Optimization: switched to using pre-tokenized data, unlike Part 1. Part 2 data now looks like:

File Data	File Size
0	5
.	2
1	3
2	1

New iterative macro:
File Manager V2.png

It seems like there is opportunity for further optimization, but this is what I have for now.

Main.yxzp

Main V2.yxzp

Gawa

part1 was simpler than I thought at first glance. I tried to create Iterative macro but I realized it's not needed after 1 hour.

For part2, I just configured the required logic into Iterative Macro. No special trick is applied and it took 11 min.

I suppose there should exist a lot of trick to speed up but cannot come up with it for now.

2024D9_gawa.yxzp

Yoshiro_Fujimori

My solution.

Spoiler

Part1

Part2 workflow
The approach for Part 2 is different from Part 1. (I tried, but could not find the way.)
Not sophisticated (it took 48:47 minutes), but it worked anyway.

Part2 macro

Part1.yxmd

Part2.yxzp

DaisukeTsuchiya

P2 was tough Challenge. My WF run in 9min for P1 and P2.

Spoiler

P1 is no macro. P2 uses itteration macro.
スクリーンショット 2024-12-10 001347.png

DT2024_Day9_Rev2.yxzp

AkimasaKajitani

Solved!

I realized that the logic was incorrect, so I had to remake the workflow that I had created several times.

Spoiler

Part1: No macro
I solved it by breaking down the file into rows, reversing the order of free space, and joining them by row position. However, free space is larger than files, the end will be a bit strange, so the key point was how to deal with that.

Part 2:
It would be a lot of work to break it down row by row, so I calculated the number of files and the free space and stuff them one by one into the free space from the back ID files. There are a total of 10,000 IDs, so I just repeated this 10,000 times. My workflow took 8 min for part1 and 2.

The point is that after the files are moved, there is continuous free space, so if you don't combine it well, you won't be able to pack the files as you intended. This is the last thing I noticed, and I ended up having to recreate the macro I created at the beginning (at first I thought it would be enough to just keep the free space in a loop, but in the end it turned out that I can't combine continuous free space unless I also have the file part).

Day09.yxzp

geoff_zath

Workflow

Spoiler

Macro

Spoiler

AoC_2024_day9.yxzp

mmontgomery

Day9. More in spoiler

Spoiler

P1 was broken into a few steps: 1 was to convert the disk map into ID number and order. 2 was to move the file blocks and I did it without a macro. I wasn't sure if ID number 10 was treated as a number in one slot or broken over two slots. Fortunately I was able to figure out that answer quick

P2 was to build the macro to make sure an entire block could move left if the last item of the block could move over. Brute forced it and it took 45 minutes

Day 9.yxzp

DataNath

Finally finished up with day 9! Part 1 was fine. Explored a few different routes for part 2 and got there after a bit of nonsense.

Spoiler

peter_gb

Possibly my favourite ever runtime and now favourite number I've generated in Alteryx.

Felt good to solve that one, I knew my logic was accurate yet inefficient so just hoped for the best and let it do its thing.
Keen to look at all these magical <1hr solutions!

Six trillion two hundred thirty-seven billion seventy-five million forty-one thousand four hundred eighty-nine
6,237,075,041,489: what a number!

Spoiler

AoC_9.yxmc

AoC_9.yxmd

JeffF

Spoiler

Part 1 macro

Part 2 macro

Day09-JeffF.yxzp

Tokimatsu

Done.

Spoiler

AoC2024_Day09.yxzp

Hub119

Finally getting around to posting this one. Not too bad once you can decipher what is being asked. I definitely went down quite the long rabbit hole on part 2 before rereading the question and realizing the ask was actually a bit simpler than what I was initially attempting to do.

Spoiler

AoC 2024 Day 9.yxzp

ntakeda

My solution.

Spoiler

AoC2024-part9.yxzp

leozhang2work

Spoiler

Part 2 the shift position of 2 caught me out a bit, took a while to handle the updated space.

day 9.yxzp

leozhang2work

@mmontgomery That's quite a list of tools 😁

Samantha_Jayne

Trying to get back in the game! Also upgraded to 24.2, welcome Dark Mode 😁

Spoiler

Samantha_Clifton_Day9.yxzp

ccanterman

Solved! I really enjoyed this one. My Part 2 initially ran in 7:20 and then I spent an hour messing around with it and got it to run in 7:10! 😂

Part 1

Spoiler

I split everything into its own cell, threw a Record ID on, and then Joined by Position.

What initially happened was that I wasn't factoring in when a Record ID was already ahead of a space - I threw on a Count of the non-nulls and then only looked at blanks that came before that number (since if it was after, I didn't care about them)

Part 2 v1

Spoiler

For Part 2, I went back to the data before everything was split. I threw it into an Iterative Macro where I checked if I could move it forward. One of the things that I did was if there were multiple blank spots with 3 spaces, I took the one with the lowest ID, since that's where the disk would move. I think that helped with my run time, compared to what other people's runs took. After each 1 was checked and either moved or left alone, I set my field Checked? to 1.

I repeated for 10k iterations without any intelligent stop condition - just find how many disks I have, and keep repeating, starting from the very last disk and working to the very first disk.

Total run time = 7:20

Macro v1.png

Part 2 v2

Spoiler

As Part 2 was running, I got to thinking about how I knew my disk #2 wasn't going to move because there was no space. I didn't need to wait until check 9,999 to determine that.

I added a Standard macro inside the Iterative macro. After the Iterative macro is mostly done, I take the lowest Disk ("RecordID") that hadn't been checked yet and saw if there was enough room for it to move up. If the answer was no, I would change the "Checked?" field to 1, which would mean that I don't need to run it through iteration 9,999 anymore. Then, the next iteration would check the next lowest Disk.

If the answer was yes (there is space), I left it alone and rechecked after the next iteration.

I wasn't brave enough to check multiple Disks at the same time. That probably would've helped materially reduce my run time.

I also modified the Iterative macro to have an intelligent stop condition which was based on when all the Disks have Checked? = 1

Run time = 7:10 and Iterations = 5,075... 10 whole seconds faster!

Macro v2.png

Day 9.yxzp

Goddenra

Part 2 of this bugged me for days, but finally got there once I found the final corner case. Really enjoyed this one!

Day09.yxzp

estherb47

This one was a fun exercise, though I'm unhappy with my runtime for Part 2, and wonder if it can be done without a macro.... Would be a heck of a multi-row formula, and I'll noodle on that for a while.

Took a long time to figure out part 2, but once I did, I got the correct answer on the first try. Which is good because the run time was a very long 2 hours 8 minutes.

Happy Holidays to all!!

Day9.yxzp

kelsey_kincaid

This is a bit of a blunt approach, but interesting problem.

Spoiler

Part 1 I solved with a sort & re-sort of the data and it was relatively straightforward once I landed on that approach
Part 2 was more complicated, and I got hung up on how to best handle free space. I had some edge cases in my data that gave me trouble, but eventually got there. May need to go back and wrap my head around all of those.

2024_Day9_Alteryx_KelseyKincaid.yxzp

lwolfie

I will continue for the second star later if I have time.

Day9.yxzp

Pilsner

Again, only got one part done on this one. Here's my solution:

Spoiler

Day 9.yxzp