Spatial Process - cut 2nd from 1st leaves too much area

[RickPack]

I am using the Spatial Process tool ("Cut 2nd from 1st") to remove the trade areas from one store list from a different, much larger store list but the spaces (trade areas) remaining are more and larger than expected. I can get the expected results by using a very small subset of the smaller store list, and activating Eliminate Overlap in the Trade Area tool's two uses (one for each store list). However, I need to use all of the data.

Does anyone know a solution? 

It looks like the Spatial Process tool's "Cut 2nd from 1st" is only sometimes removing spaces.

Notice how at the bottom-right, the southeast area of Florida remains in the "Cut 2nd from 1st" region even though it is present in the 2nd.

I regret that I cannot provide the data.

Cut 2nd from 1st

2nd [removed by Spatial Process but notice that at the southeast of the U.S., much of the Los Angeles area appears to not be removed)

1st (large store list)

[Ben_H]

Hi @RickPack 

I think it would be simplest for you to summarise the trade areas first into a single spatial object each for List 1 and List 2.

Then perform your spatial process.

If you then need to know which stores remain in the area with no overlap just spatial match the centroids from the appropriate list.

It looks on the face of it like you're pairing the different store trade areas somehow so all of the spatial processes are slightly different. I may be wrong though. It would help if you could attach a workflow with some dummy data.

Regards,

Ben

[RickPack]

You may have just pointed out a data structure element unknown to me.
Are you saying that a "single spatial object" can be noncontiguous spaces that still get processed by Alteryx in such a way as to be fully represented in subsequent tool usage (like Spatial Match)?

Yes, I paired the trade area with a spatial match tool thinking that the stores in the small list would likely always intersect the radius of a large list.

If it helps, I am looking for the population in the post-removal area (U.S. Census Block Groups - Allocate Input tool). I have successfully used the Allocate Input tool for each store list.
The work discussed in this Community question is for validation.

Thanks, Ben!

[Ben_H · Accepted solution]

Hi Rick,

Yes that's right, I've attached a simple example of how I would do a similar task.

It's UK data but it works the same.

I've taken a set of UK county centroids, split them into a large and small group to stand in for your stores.

Then I've summarised the trade areas together so each set is a single spatial object. Then appended the 2 groups together.

I've then cut 2 from 1. I also clipped it to a UK shape file for purely aesthetic reasons!

You can then just use the allocate tool on the remaining spatial obj to get the TOTAL population, A single spatial object can have multiple different "parts" but be counted as 1 object.

Regards,

Ben

EDIT - attached the wrong file before