I have two feature classes that I want to create a table from doing a join and update on. However, I don't want to alter the original feature classes but instead would like to create layer files to perform the join and update processes. I drafted a python script that creates the final output table of these operations just fine but noticed it is also joining the join features (streets) to the original target feature class (GIM) and updating the table. This may cause problems as these target feature classes need to retain their schemas. I thought making layer files and running the geoprocesses on them would prevent this. Here is my code:

import arcpy

# deignate output location and input variables    
outPath = raw_input("Enter output workspace gdb: ")    
GIM = raw_input("Enter geocoding information manager: ")    
streets = raw_input("Enter last year's street file: ")    
outName = raw_input("Enter an output name: ")

# initiate geoprocesses    
    print "create layer files to run geoprocesses and select by location queries on"
    arcpy.MakeFeatureLayer_management(GIM, "gimLyr")

    arcpy.MakeFeatureLayer_management(streets, "strLyr")
    print "made lyr files"

    print "Select by location on streets for data reduction"

    # Select by location args
    distance = 40
    unit = str(distance) + " Feet"

    arcpy.SelectLayerByLocation_management("strLyr", "WITHIN_A_DISTANCE", "gimLyr", unit)
    print "Select by location succeeded"

    arcpy.MakeFeatureLayer_management("strLyr", "targetStrLyr")

    print "Add fields for street name comparison"
    fieldLength = 10

    arcpy.AddField_management("gimLyr", "Cust1stStWord","TEXT",field_length=fieldLength)

    arcpy.AddField_management("gimLyr", "Street1stStWord","TEXT",field_length=fieldLength)

    joinFields = ["LINK_ID","ST_NM_BASE"]

    print "Join layers"
    arcpy.JoinField_management("gimLyr", "Ref_ID", "targetStrLyr", "LINK_ID", joinFields)

    print "Make new layer file from join"

    f1 = "Address"    
    f2 = "Cust1stStWord"    
    f3 = "ST_NM_BASE"    
    f4 = "Street1stStWord"    
    f5 = "Ref_ID"    
    v1 = 10000    
    numbers = ('1ST','2ND','3RD','4TH','5TH','6TH','7TH','8TH','9TH','10TH')         
    alpha_num = zip(alphas,numbers)

    print "Running cursors"
    with arcpy.da.UpdateCursor("gimLyrJoin",[f1, f2, f3, f4]) as cursor:
        for row in cursor:
            # fix null values
            if row[2] == None:
                row[2] = " "
            elif len(row[0].split(" ")[1]) == 1 and row[0].split(" ")[2] != "ST":
                row[1] = row[0].split(" ")[2][0:3]
                # ST_NM_BASE seems to not include prefix so okay to 0 index
                row[3] = row[2].split(" ")[0][0:3]
            # now look for street names that start with SAINT in customer records and change them to ST
            elif row[0].split(" ")[1] == "SAINT":
                row[1] = "ST "
                row[3] = row[2].split(" ")[0][0:3]
            #Change remaining records to first three letters of street name
                row[1] = row[0].split(" ")[1][0:3]
                row[3] = row[2].split(" ")[0][0:3]

    # fix street name discrepancies between alpha and numeric entries
    with arcpy.da.UpdateCursor("gimLyrJoin",[f1, f2]) as cursor:
        for row in cursor:
            f1_elements = row[0].split(" ")[0:]
            for x in alpha_num:
                if x[0] in f1_elements:
                    row[1] = x[1]

    q2 = """{0} <> {1} OR {2} < {3}""".format(f2,f4,f5,v1)

    print "Making final table"
    arcpy.TableToTable_conversion("gimLyrJoin", outPath, outName, q2)

    print "succeeded!"
except arcpy.ExecuteError:


Is there something I'm missing that would cause both the layer file and the original feature class to be updated?

  • A layer file is just a pointer to data that is on disk. You use them to define symbology, field display, scale, etc. If you want something temporary, copy the feature classes to a scratch workspace (in_memory is nice for this) and do your work on that. – Paul Jan 13 '17 at 20:04
  • So arcpy.MakeFeatureLayer is not making an in_memory file? – geoJshaun Jan 13 '17 at 20:49
  • It's just creating a temporary layer (would become .lyr(x) if you saved it to disk) which you can use for selection and such. The data is still on disk. MFL and the like is really useful when you want to run a GP tool on a subset of data. Instead of creating a new FC, just select the features you want and pass that to the GP tool. – Paul Jan 13 '17 at 21:01
  • Thanks Paul. I'm still confused as to why it's changing my input data. I get that the layer file saved to disk is going to be processed by why is it also processing the feature class it was made from? – geoJshaun Jan 13 '17 at 21:04
  • You're missing the point of what a layer file is. It's only a few KB, right? All that data is still stored in the dataset. The lyr file stores no tabular data. – Paul Jan 13 '17 at 21:07