Arcmap 10.2.2 crashes when I perform a spatial join

3136
8
10-08-2015 10:58 AM
YousefAbdalAal1
New Contributor

Greetings,

I am trying to perform a spatial join between a polygon shapefile of about 27 records representing weather stations area cover (created from a buffer of a point shapefile) and a polygon shapefile representing farm areas with a field (of type double) having the area of each record for a total of 92302 records (created from raster conversion)

I am just interested in the total area of farms contained inside each of the buffer polygons of the weather stations, so I used a spatial join with buffered weather stations as the "target layer" and the other farm area polygon shapefile as the "join layer" through a "JOIN_ONE_TO_ONE" procedure, I just included the area field from the farm areas polygon with a "merge rule of sum" and removed the other fields of this join layer from the "Field Map of Join Features window". I chose the "completely contains" as the match option.

I tried putting all the input layers inside a geodatabase and saved the output inside the same geodatabase but it was the same problem,

I get an error If I keep the background geoprocessing option enabled, and ArcMap crashes if I disable background geoprocessing.

When disabling the background geoprocessing it always crash when it is performing "Final selection step".

I should mention that the buffer polygon shapefile for the weather stations have some overlapping/intersecting records. I am not sure but this isn't supposed to be the cause of this problem.

I don't think the problem was a result of too many records to process problem, because I created two small shapefiles to mimic the process and test it but got the same error.

Attached is a zipped file for the geodatabase containing the files I used.

My operating system is windows 8.1, 64-bit, have 6gb rams, intel core i5 -4200m @ 2.5ghz, using a student version ArcMap 10.2.2

0 Kudos
8 Replies
DanPatterson_Retired
MVP Emeritus

You need to check and report your coordinate systems to ensure they are the same.  Also, make sure you try with singlepart shapes just in case that is a problem

0 Kudos
YousefAbdalAal1
New Contributor

Greetings,

Both of the shapefiles have the same coordinate system.

I am not sure what do you mean by single part shapes, do you refer to "the one to one" or "one to many" option in spatial join tool interface?

I Still cant perform this analysis, always crashes.

0 Kudos
DanPatterson_Retired
MVP Emeritus

A multipart feature is one that has two or more separated geometry pieces but share a single record in a table as shown here Multipart To Singlepart—Help | ArcGIS for Desktop

that tool converts those to a different representation of the data.​ You don't indicate whether the join works on a subset of the data.  This could be investigated by either tiling your data and  running the test on the pieces or subsetting by query or selection to reduce the number of inputs and hence memory consumption. 

As a final check  I assume that the data passes the check geometry Check Geometry—Help | ArcGIS for Desktop   and repair geometry conditions (if needed) Repair Geometry—Help | ArcGIS for Desktop

0 Kudos
NeilAyres
MVP Alum

Had a look at your data and spatially joined poly_area to buff_15 with the sum option.

Took about 2mins. My version 10.3.1 on Win7 64bit.

0 Kudos
NeilAyres
MVP Alum

But, better do some tests here on how the spatial join handles the overlapping bits.

Not sure if it will "count twice" these farms / fields in the overlapping portions.

0 Kudos
NeilAyres
MVP Alum

Just did a quick test.

Yes, the spatial join does correctly handle the underlying features. It counts them twice when summing into the larger overlapping polygons.

YousefAbdalAal1
New Contributor

First of all Thanks a lot for giving time to try working things out.

I will try to operate the function from another PC and see if I have the issue was from my labtop or OS.

by counting twice you mean if there is a field that exists in 2 different buffer zones such that it is taken into account for the final area computed at each of the buffer zone?

Thanks again

0 Kudos
NeilAyres
MVP Alum

Yes,

the underlying small polygon is summed into both of the larger overlapping areas. Which is just what you want in this analysis.