Generate spatial weight matrix from file in ArcGIS

Question

I am trying to create a spatial weight matrix from an existing file in ArcGIS. I need to use another file, as opposed to generate the weights in Arc, because I define non-standard weights. The documentation on how to do so is a little scarce, so I am turning for advice as my attempts have failed so far. All the files can be found in the zipfile downloadable through this link in Fileshare.

My locations are a set of points in the shapefile "pointlocations.shp". My ascii weights table is "weight_table.csv". The inuque id field is "codice71". I used the tool from the toolbox "generate spatial weight matrix" in the following way (I copy-pasted the python snippet here):

arcpy.GenerateSpatialWeightsMatrix_stats(Input_Feature_Class="PATH/pointlocations.shp", Unique_ID_Field="codice71", Output_Spatial_Weights_Matrix_File="PATH/spatweight4.swm", Conceptualization_of_Spatial_Relationships="CONVERT_TABLE", Distance_Method="EUCLIDEAN", Exponent="1", Threshold_Distance="", Number_of_Neighbors="", Row_Standardization="ROW_STANDARDIZATION", Input_Table="PATH/weight_table.csv", Date_Time_Field="", Date_Time_Interval_Type="", Date_Time_Interval_Value="")

When I do this, I run into the following error:

Runtime error  Traceback (most recent call last):   File "<string>", line 1, in <module>   File "c:\program files (x86)\arcgis\desktop10.4\arcpy\arcpy\stats.py", line 1423, in GenerateSpatialWeightsMatrix     raise e ExecuteError:  Traceback (most recent call last):   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 848, in <module>     setupWeights()   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 164, in setupWeights     rowStandard = rowStandard)   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 795, in table2SWM     allMaster.remove(lastID) ValueError: list.remove(x): x not in list  Failed to execute (GenerateSpatialWeightsMatrix).

I have no clue of what is going on, and I have found little help to find a better approach using the online help in: Generate Spatial Weight and Modelling Spatial Relationships

I was trying to check whether the formatting of my table was wrong, so I tried to have one that looked closer to the one described on the page Modelling Spatial Relationships. In that attempt, instead of having the csv file "weight_table" with header codice71, nid, weight, I had the txt tab-separated file weight_table2 with just the header codice71. Doing so, I get the error:

Runtime error  Traceback (most recent call last):   File "<string>", line 1, in <module>   File "c:\program files (x86)\arcgis\desktop10.4\arcpy\arcpy\stats.py", line 1423, in GenerateSpatialWeightsMatrix     raise e ExecuteError:  Traceback (most recent call last):   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 848, in <module>     setupWeights()   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 164, in setupWeights     rowStandard = rowStandard)   File "c:\program files (x86)\arcgis\desktop10.4\ArcToolbox\Scripts\Weights.py", line 772, in table2SWM     for row in rows: RuntimeError: A column was specified that does not exist.  Failed to execute (GenerateSpatialWeightsMatrix). 

Any help on what I am doing wrong? I am also open to using other software (R or Python) to do this, bearing in mind that my goal at the end is to generate a hot-spot detection map (Local Moran stats with pseudo p-values).

Answer 1

It seems as though the table must be formatted with the unique ID field in order (or at least grouped). In the Weights.py script which is throwing the error there is a function that reads the table and checks the value against a list. Every time the tool processes a new unique ID it removes it from the list. If the IDs are not grouped it will fail when trying to remove that ID because it has already been removed. I don't see this in any of the documentation so it may be a bug.