# Weekly Challenge

Solve the challenge, share your solution and summit the ranks of our Community!
New content is available in Academy! You may need to clear your browser cache for an optimal viewing experience

## Challenge #161: Triangles, Triangles, Triangles

Highlighted
Alteryx

Here's my solution. Had a lot of fun playing with various sorts on the final data set, trying to figure out how to match the image (area, min X, min Y in ascending and descending order). Came up with some interesting patterns, especially the last two which are sorted on area in ascending and descending order. But I couldn't find the sort order that matched the image.

Also learned a few interesting things looking at the solution:

Spoiler
I didn't realize you could do comparisons on polygons. I thought that Alteryx was comparing the underlying GeoJSON so the order of the points would matter. Apparently it's smarter than that, which makes the problem a lot simpler. Another thing that makes my workflow longer is that I tried to stick to what I thought was the spirit of the exercise and use spatial tools instead of formulas whenever possible.

Highlighted
Asteroid
Spoiler
Alteryx Alumni (Retired)

Highlighted
Meteoroid
Spoiler

My associate @CharlieS came up with this one!

Highlighted
Asteroid

Highlighted
Asteroid

Definitely not an optimal solution, but, hey, it works!

Spoiler
Highlighted
Meteor

Highlighted
Asteroid

The extra 21 objects (537-516) seem to be true triangles - the nulls are lines. (I added collinear test records to check). So does that mean if 3 collinear points are used in st_createpolygon that it won't produce a spatial object?

Highlighted
Alteryx Certified Partner

@garthn555 wrote:

The extra 21 objects (537-516) seem to be true triangles - the nulls are lines. (I added collinear test records to check). So does that mean if 3 collinear points are used in st_createpolygon that it won't produce a spatial object?

Yes, 3 perfectly collinear spatial points will yield "ST_CreatePolygon: Failed to form valid polygon."

That being said, the issue with this scenario is introduced by the Make Grid tool and the issues of plotting a flat plane to a curved surface. The grid points are not technically collinear, but practically they are, and therein lies the issue.

Highlighted
Quasar

I was stuck for the longest time at 540 triangles before I peeked at the solution to see that the area tolerance was set to > .1 sq miles.

My solution:

Spoiler