Ill try my best to clearly explain what I am trying to get out of this. Been working in a non for profit in Chicago for about a year I am currently working with a number of different shapfiles that deal with protected lands; PAD-US, DNR, Conservation Easements, etc.. The goal of the project is to create a master data set that combines all the individual protected lands datasets in order to have a comprehensive understanding of all of the protected areas in my region. We have started to look into how to best go about dealing with overlapping polygons and want to see if anyone else out there has any ideas or help with this.
1. The image that is attached has 4 of the data sets on it, in the order that we want ( ie if ly1 is overlapping 2,3,4 we want to take ly1 and discard 2,3,4.) We are currently struggling with the proper way to identify these overlaps. Centroid was our first idea, (if polygons centroid of ly2 is within polygon of ly1 then take layer1, discard ly2) How ever what happens with overlapping polygons where their centroids are not within the target layer polygon(example in lower right corner of image)? Is there a way to identify the percent overlap of the polygons as a way for us to check which dataset we should use?
2. What is the best way to deal with the attribute tables? Our methodology so far; take each dataset, exploded polygons, so that each polygon is its own, then give them each their own unique source_id, so we at least know what dataset each polygon came from. But then how does the output attribute table work? Can we have an output that selects the target layer along with an output of the discarded polygons so again we can check the accuracy of the model.
Any information would be greatly appreciated, let me know if I am not being clear, thanks in advanced!
It would sound like you would want to Union the data sets. That should give you a single, planer, layer that maintains all the original boundaries and all the original attributes for each 'piece of the puzzle'. If there are a lot of attributes in each original layer, union either with ONLY_FID or strip out everything from the inputs except a more informative unique ID, and use ALL. Then you can link back to the original shape file to access the other attributes. Without an Advanced lic. you will have to Union cumulatively, adding each layer one at a time: with Advanced, you can just throw them all in.
It is a bit confusing when you write about 'taking' and 'discarding'. Union will keep everything, but will tell you that a particular polygon will belong to layers 1, 3 5,6, and 7, while the one next to it belongs to 1 and 7 (for example)
Thanks for the replies back, love forums that have friendly knowledgeable users. Thanks for the information mdenil, we've been talking about that idea as well. Ill take a look at the spaghetti and meatballs article, nice name :), thanks for pointing me in the right direction. Ill report back in a bit tell you how we all went about solving this issue.