Iteratively unioning pairs of shapefiles

Question

A simple (I think) Model Builder question. I have 25 field sites, for each of which I have mapped land use as separate layers for different years. In one folder I have veg layers for 1994, named tara{site number}veg94 (eg., tara01veg94.shp, tara02veg94.shp, through to tara25veg94.shp). In the other folder I have the same for 2007, similarly named by site number (tara01veg07.shp, tara02veg07.shp ... tara25veg07.shp).

I am wanting to union the layers for each site for successive years together using a model, for example tara01veg94.shp unioned with tara01veg07.shp to create tara01veg9407 and so on. I presume that I need to use an iterator to specify the site number for each run of the union function and then insert the %value% created by the iterator into the input and output file names. The problem I'm having is that I need to iterate the same site number for two input files simultaneously, so I figured the "for" iterator was the one to use, but the union function doesn't let me put %Value% into the input fields (though it accepts it in the output field ok). Any help would be greatly appreciated... thanks.

score 2 · Answer 1 · edited May 23 '17 at 12:39

2

I would use Python instead of ModelBuilder for something like this.

The gist of it would be to do something like this:

Read the contents of the first folder into a list using ListFeatureClasses
Do the same for the second folder
zip() the two lists into a dictionary: Example
Iterate over the dictionary's key/value pairs (Example) and perform the Union for each pair

edited May 23 '17 at 12:39

Community

1

answered Mar 13 '13 at 05:58

blah238

35,793
7
94
195

score 2 · Answer 2 · edited Mar 14 '13 at 02:52

2

Thanks for recommending Python. Having never used it before, I've had to take a crash course in Python programming. I stumbled when it came to iterating through a dictionary as you suggested, so I tried another approach -- which seems to work. I'll paste it here for anyone else if it is of use.

import arcpy
arcpy.env.workspace = "c:/Temp/TRC_processing/processing_auto"
for i in range(1,25):
    veg2012 = "c:/Temp/TRC_processing/2012_layers/tara" + str(i) + "v12_Dissolve.shp"
    veg1994 = "c:/Temp/TRC_processing/1994_layers/tara" + str(i) + "v94_Dissolve.shp"
    vegchange9412 = "tara" + str(i) + "v9412_change.shp" 
    arcpy.Union_analysis([veg2012, veg1994], vegchange9412, "ALL", "", "NO_GAPS")

edited Mar 14 '13 at 02:52

blah238

35,793
7
94
195

answered Mar 14 '13 at 01:02

harley

51
2

An easy way to iterate over the key/value pairs is to use tuple unpacking and dict.items() or dict.iteritems(), e.g. for k, v in {"a":1, "b":2}.iteritems(): print k, v – blah238 Mar 14 '13 at 01:55
You might also be interested in this question: What are some resources for learning Python and arcpy? – blah238 Mar 14 '13 at 02:54

score 0 · Answer 3 · answered Mar 14 '13 at 05:16

Unless you must use different projections among your sites or you have polygons that must overlap, or have another valid reason not to combine all your sites, Merge all the data for each year so you can then efficiently perform one union between each pair of years.

Prior to merging ensure that each site-veg layer has a field that identifies the site ID. The ACCRU Tools 1.2.1 Merge With Filename has an automated way to add/calculate this in, if not already present in the tables of the files to be merged.

merge all files in folder 1994 (tara01veg94.shp, tara02veg94.shp, ... tara25veg94.shp) = veg94.shp
merge all files in folder 2007 (tara01veg07.shp, tara02veg07.shp, ... tara25veg07.shp) = veg07.shp
and then union veg94.shp and veg07.shp to get a layer that you can calculate an updated area field and perform Summary Statistics using the site ID field as one of the case fields.

You will likely want all combinations in one final table for further statistical analysis, so doing the merges up front means you don't have to iterate all geoprocessing, including adding/calculating fields, etc.

If your sites overlap, then you will need to union your veg layers and probably do some clean up in the attribute table and also maintain a separate layer for your site polygons (these will need that unique site ID field) to be used to intersect with the unioned pair of years. I can provide more explanation of all this if needed (if no overlap among sites, then ignore this last paragraph).

Thank you ccn. I also had this approach suggested to me by a colleague here. There's no overlap between the sites so this approach would be useful (and more efficient) too. Have added your comments to my notes. — harley, Mar 17 '13 at 22:07

Iteratively unioning pairs of shapefiles

3 Answers3