Splitting shapefile based on part of field in attribute table using ArcPy?

Question

I'm new to Python and I need a script that can split my huge shapefile based on a certain attribute.

SO what I have is a big shapefile (over 3M records) with a field named "NISCODE". In this code are values ranging from 11001 to 73109. But only the first number of the value is important to me, that's the base on which the shapefile needs to be split on. So I need 7 shapefiles; everything starting with 1, with 2,...

I know I'll probably need a for loop, but I don't know how to put it in my code correctly. Here is what I have so far, this was just a test to see if I could already make a selection for the 1st part (everything starting with 1...), but already that fails.

# Import system modules
import arcpy
from arcpy import env
env.overwriteOutput="True"

# Set workspace
# setting the environment
workspacedir= r"C:\..."
env.workspace= workspacedir

shp = r"C:\....\Adp.shp"

r=arcpy.Exists("Adp.shp")
print r
### Make a layer from the feature class
arcpy.MakeFeatureLayer_management("Adp.shp", "Vl_lyr")
##
# Wselect only those that have id starting with 1
arcpy.SelectLayerByAttribute_management("Vl_lyr", "NEW_SELECTION", ' "NISCODE" = 1* ')
##
# Write the selected features to a new featureclass
arcpy.CopyFeatures_management("Vl_lyr", "Vlaanderen_1.shp")

EDIT:

The new code looks now like this:`

# Import system modules
import arcpy
from arcpy import env
env.overwriteOutput="True"

# Set workspace
# setting the environment
workspacedir= r"C:\..."
env.workspace= workspacedir

shp = r"C:\...\Adp.shp"

arcpy.AddField_management(shp, "newNISCODE", "TEXT")

arcpy.CalculateField_management(shp, "newNISCODE", '"newNISCODE" = NISCODE[0]', "PYTHON")

list = []
with arcpy.da.SearchCursor(fc, "newNICODE") as cursor:
    for row in cursor:
        list.append(row)
num_list = sorted(set(list))

for i in range(1, len(num_list)):
    arcpy.Select_analysis(shp, "{}_shp".format(i), 'newNICODE = "{}"'.format(i))`

but the code fails on the Calculatefield step:

Traceback (most recent call last):
  File "C:\...\SplitShape.py", line 15, in <module>
    arcpy.CalculateField_management(shp, "newNISCODE", '"newNISCODE" = NISCODE[0]', "PYTHON")
  File "C:\Program Files (x86)\ArcGIS\Desktop10.5\ArcPy\arcpy\management.py", line 3663, in CalculateField
    raise e
ExecuteError: ERROR 000539: SyntaxError: invalid syntax (<expression>, line 1)
Failed to execute (CalculateField).

Answer 1

you don't need to compute a new field. In your code, the first problem is that you handle the NISCODE field as a numeric field, but it can be seen that it is actually a text field (text are left aligned, number are aligned right). Therefore you need ' ' around the value. Second, if you test with a wildcard you must use LIKE and not =. Third, the SQL wildcard character is % (or _ for a single character). So instead of:

arcpy.SelectLayerByAttribute_management("Vl_lyr", "NEW_SELECTION", ' "NISCODE" = 1* ')

it should be

arcpy.SelectLayerByAttribute_management("Vl_lyr", "NEW_SELECTION", """ "NISCODE" LIKE '1%' """) 

Alternatively, you could use some SQL syntax to take only the first character

arcpy.SelectLayerByAttribute_management("Vl_lyr", "NEW_SELECTION", """ LEFT("NISCODE",1) = '1' """) 

then you can make a loop:

mylist = []
with arcpy.da.SearchCursor(shp, "NISCODE") as cursor:
    for row in cursor:
        mylist.append(row[0])
num_list = list(set(mylist))
for i in num_list:
    arcpy.Select_analysis(shp, shp[:-4] + i +".shp", """ Left("NISCODE",1) = '{}'""".format(i))

Note that shp fieldnames must be written between " " , but it would not be the case in gdb.

As a remark, your calculate field expression is not working because 1) you mustn't write the target fiel name in the expression and 2) python field names should be written with ! !.

Answer 2

You could add a new field, and calculate that field based on the field you need. Using python, we can recalculate it to only contain the first digit. It looks like the NISCODE is text so try this:

arcpy.AddField_management(shp, "newNISCODE", "TEXT")

arcpy.CalculateField_management(shp, "newNISCODE", '!NISCODE![0]', "PYTHON")

list = []
with arcpy.da.SearchCursor(shp, "newNICODE") as cursor:
    for row in cursor:
        list.append(row)
new_list = sorted(set(list))
num_list = [item for t in new_list for item in t]

for codenum in num_list:
    arcpy.Select_analysis(shp, 'shp_' + codenum + '.shp',  'newNISCODE' = codenum)