Advent of Code 2025 Day 9 (BaseA Style)

AlteryxCommunityTeam

Discussion thread for day 9 of the Advent of Code - https://adventofcode.com/2025/day/9

Advent of Code

Accepted answers

All comments

Gawa

Alteryx spatial analysis really do great job!

Spoiler

Part1
Just calc area for all rectangular patterns and find max. It's simple task. Offset +1 is needed.

(ABS([X]-[Source_X])+1)*(ABS([Y]-[Source_Y])+1)&#8203;

Part2
Overall combined polygon is needed. My approach was to create new 3 points for each coordinate so it forms square, and trace each point individually to create sequence polygon. Then we will have 4 individual sequence polygon, so combine them into one. This is overall polygon map.

Here is how it looks like with my dataset.

Next, create all rectangular patterns as spatial polygon object as well, and find rectangular that are contained within overall polygon by the Spatial Match tool.

Once you get the target rectangular, then you just have to find the one with max area.

2025D9_gawa.yxzp

CoG

@gawa, great explanation! This was a fun problem except for a misunderstanding that cost me a lot of time...

Spoiler

The Spatial Match Tool is not perfect. There can be overhangs from this tool. I had to include a Spatial Process Tool to be able to get the actual answer.

Hint: Graph the points to identify a pattern that is easier to search!

Workflow.yxmd

Tokimatsu

Spatial tool is good.

Spoiler

Logic is very simple. I created the box of each two points. When box forms just line, I used st_createline. Then processed.

スクリーンショット 2025-12-09 154736.png

AoC2025Day09_toki.yxzp

Hub119

So it turns out (as we all know) where you place parentheses in an equation matters... strangely enough mistyping my area equation in part 1 didn't throw off my solution... it just became the bane of my existence in part 2 as I spent an hour plus looking in the wrong places to fix the problem.

Spoiler

One thing to note, as @CoG also called out, the spatial match tool has some issues with this problem, so in all my checking for fixes I ultimately got rid of it and just went straight to the spatial processing.
AoC 2025 D9 Pic.png

AoC 2025 Day 9 - Movie Theater.yxzp

Raj

Realized I need to spend more time with spatial tools

2025 D9.yxmd

Qiu

My brain hurts. @Hub119 Thank you for the tip about the Spatial Match tool.

Spoiler

AoC 2025 Day 09.yxzp

DanFlint

It's been ages since I've had a good reason to use spatial tools, I really enjoyed this one!

Spoiler

No need for spatial in part 1, just calculate side lengths.
Part 2 builds a shape which is only green and red tiles, then builds each possible rectangle. Spatial match to determine which rectangles are entirely within red and green tiles, then calculate the area in the same way as part 1.

Day9_DF.yxzp

martinson

I love a chance to use the Spatial Tools

Spoiler

AncientPandaman

wastes 3hrs for spatial tools. turn out I got chance to find the answer if I not confuse cut1_from2 and cut2_from1.

Day 09.yxmd

DaisukeTsuchiya

I spent two hours debugging issue.

Spoiler

In P1, I simply generated all combinations and calculated the area of each rectangle.

For P2, I initially struggled with sorting the 496 input points, not realizing they were already ordered sequentially to form a continuous polygon. This oversight cost me significant time in constructing the polygon. Moreover, the Spatial Match Tool produced inaccurate results for the largest area, forcing me to visually inspect and confirm the correct answer. To top it off, I spent about another two hours only to discover that my area calculation formula was incorrect.

スクリーンショット 2025-12-09 195609.jpg

DT2025_Day9_Rev2.yxmd

DataNath

Pretty nice challenge today considering we're on Day 9! That is... (spoiler)

Spoiler

If you're not a complete idiot that spends 1.5 hours debugging spatial processing just to notice that your area calc isn't correct.

Shoutout @Hub119 for the Spatial Process prompt - didn't click on to what he was ranting about in the WhatsApp chat until I ran into that brick wall myself!

Day 9 - Cleansed.yxmd

AkimasaKajitani

Part of the problem was difficult to understand, but it seems I took a long detour because I didn't read the problem carefully.

Spoiler

In order to create the outer polygon, I went to the trouble of finding combinations, rearranging the lines, and working hard to create the polygon, but all I needed to do was create it in the original order.

My 1st workflow.

Spoiler

Macro

After reading community.

Spoiler

Day09_2.yxzp

Day09.yxzp

Goddenra

Was determined to avoid using the Spacial tools, went down a complete rabbithole by building out a grid using formulas, realised the maths would never allow that to work, gave up and force myself to use Spacial.
Really wish I had, as it took a fraction of the time. And isn't that what AoC is supposed to be about, using things you don't typically use.

Spoiler

Thanks to those noting that Spacial Match didn't generate correct result, lost a lot of time on that, before taking Spacial Process approach.

Couldn't initially figure out how to generate the rectangle from the two points, before I learned about the bounded rectangle formula (thanks Google!).

Day9-Take2.yxmd

mmontgomery

Day 9

Spoiler

Thank god for spatial tools!
P1: Rather easy brute force by appending all points to each other then taking area of min/max points on rows/columns
P2: Took a bit on approach but landed on spatial. First I created a Sequence Polygon of all the points (don't forget to add the initial point to the end to close the loop). Then I created a bounding rectangle of point a, point b then point a again. From there, I did a spatial match where target within universe to see which bounding rectangles fit into the sequence polygon universe. From there, take the max area and bam, there you have it!

Day9.yxzp

ScottLewis

Went around in circles trying to get a math and macros solution before recalling that this was a spatial tool first.

The math solution produced some delightful outputs for the test data and then overflowed the temp directory on my hard drive with the actual data.

Came around for some help as spatial is something I don't normally have reason to do.

Things that tripped me up:

Spoiler

Dividing points by 1000000 if you use formulas to create them rather than the tools. I don't understand it, but I know it to be true.

That you want the Nulls out of spatial processing, not the non-nulls. The Nulls are the ones where nothing was left over when you cut the polygon of red points out of the rectangle. I remember something on one of these discussions where a wiser spatial user than me made the point that we do ourselves no favors by calling things the Universe when it's really just another polygon.

Trying to use spatial match instead of spatial processing. It is interesting that one of the tools is concrete and the other is fuzzy.

AOCDay9ScottLewis.yxmd

lwolfie

I kept changing too many things on the spatial tools at one time. So it kept not working.

Day9 law.yxmd

WirkKarl

These Advent of Code challenges are always a fun way to test different approaches in Alteryx. Looking forward to seeing how others tackle it and comparing solutions.

Stosh

Day 9 is officially in the books! This was a really cool challenge, and yet another opportunity to get some practice with some seldom-used tools. We are now 3/4 done with AoC!

My Solutions

Spoiler

My Reflection:

Spoiler

Part 1 was super easy. Brute force took me all the way home. I made a list of every possible combination, calculated their areas, and isolated the largest one. Honestly, I probably could have skipped the part where I remove pairs of the same point, but oh well! At least it proves I am aware of the issues involved with cartesian joining a table to itself.

Part 2 was where things heated up, as usual! I heard rumblings all day about how spatial tools are super useful here (Given the fact we are working with a grid, this makes sense) but that Spatial Match was giving people trouble. That being said, it was my original plan while I was dreaming up answers to the question so I just HAD to try it. I started with the same combination of every pair of points and began calculating the missing corner points. Once I had those, I created a square polygon for each pair to act as my Targets. Then, I sequenced the original list of points and created a sequence polygon to act as my Universe. When I got a wrong answer (too high), I was like: "Hmm, the correct one must be right on the edge or something... I know! I will add a buffer to my sequence polygon!" (You fool, they said too high, and buffer would only make that worse!!) Anyways, a few more fumbles with spatial match occurred before I finally decided to scrap it. I almost scrapped spatial analysis entirely, but I had already come so far that I forced myself to at least drag a bunch of other spatial tools onto the canvas to see if it sparks any inspirations. Lucky for me, I eventually landed on the Spatial Process tool and realized that I might be able to work the other way around. That is to say, if removing the 'Universe' from a target polygon leaves you with nothing left, that means all of the target polygon was within the green/red squares, and that the pair in question is valid for the problem. After I made that discovery, I was free and clear!

Overall, this problem was really cool, and I am glad we are entering into the days where the tools become super niche. Even more so, I am glad to be a part of the community that helps me so much along the way. See you all tomorrow!

AOC2025.9.yxzp

Erin

Since I take to Spatial tools about as well as vampires take to garlic, I only used one! And it wasn't the spatial match tool either!

Spoiler

So I only technically used one spatial tool to create a polygon, but I used a bunch of spatial functions!

Pt 1 -
Pretty similar to others, I did a cartesian join, removed dupes, calculated the area and found the max area.

Pt 2 -
Only because I had the advantage of reading WhatsApp comments about how spatial match tool was not being friendly (thanks @Hub119, your frustration saved me today!), I decided to cut my little rectangles from the "Big Spatial Obj" (I'm so technical), and then calculate the area of the cut object. I only kept where the cut object area was null or 0. For funsies, I did a within formula too - but that formula didn't always work. The cut objects with a 0 area were falsely excluded from my within formula.

2025_AoC_ErinMiller_Day_09.yxzp

Qiu

Really need to learn more about Spatial Tools. 😁

danboll_life

I was surprised that the Spatial tools didn't return the correct answer.

Spoiler

AoC2025_09_DL.yxzp

EKasminsky

Spoiler

Part 1 solved.. not sure on part 2.

Day 9 EK.yxmd

estherb47

Today the strangest thing happened to me. Part 1 solved easily with Spatial tools. And without (you can generate all of the points in an area with two generate rows tools, but that's TORTOISE levels of slow.

My spatial approach looked valid. Checked everything multiple times. All was in order. Thought perhaps I had exceeded the limit of integer spatial points, but nope. Thought I had forgotten to set data types correctly. Nope.

Before going the slow route, recreated my part 2 in a new workflow. Step for step, the same. And tada! The answer was correct. Somehow the first and second largest were inverted in the first workflow. Same exact tools, same exact order. Same formulas came up with different values for the combination of points.

I have to delete and burn my first workflow. Clearly, it's cursed.

Spoiler

Straightforward spatial analysis. Don't forget that the lat/lon are integers.

I can't upload my workflow from my work computer. Will do so when I recreate on my home laptop

mkeiffer

My Day 9 solution!

Spoiler

2025_AoC_MPK_Day_09.yxzp

Pilsner

Here's my day 9 workflow. I faced similar challenges to others with fractional overhang on some of the regions, but still loved the opportunity to use the spatial tools again!