Script/tool to Sum the Count of each "state" in each FC in a FGDB

Question

I am trying to generate a script that will scan a fgdb with feature classes all within feature datasets, and then it will output to a csv with two columns: state and count, and these will represent which state (US states) are in how many FCs.

The tool will parse the FGDB and FDS, then scan the FCs in the first FDS, but then not manage to search each subsequent FDS after the first one. It also does not write anything into the csv except the headers 'state' and 'count.'

pardon the abundance of messages and try/except stuff, I'm not very proficient (just kidding, I started from chatGPT).

import arcpy
import csv

# Set the workspace to the folder containing the feature classes
workspace = arcpy.GetParameterAsText(0)
arcpy.env.workspace = workspace
arcpy.AddMessage("workspace: " + str(workspace))

#feature_classes = arcpy.ListFeatureClasses()
feature_datasets = arcpy.ListDatasets("", "Feature")
arcpy.AddMessage(str(feature_datasets)) #outputs feature datasets properly
state_counts = {}

for fds in feature_datasets:
    arcpy.env.workspace = fds
    arcpy.AddMessage(str(fds)) #properly scans through each fds
    # Print the current feature dataset to verify if it's correctly set
    arcpy.AddMessage(f"Processing feature dataset: {fds}")
    try:
        feature_classes = arcpy.ListFeatureClasses()
        # Print the number of feature classes found in the current feature dataset
        arcpy.AddMessage(f"Number of feature classes: {len(feature_classes)}")
        arcpy.AddMessage(str(feature_classes))
        if feature_classes:
            try:
                for fc in feature_classes:
                    arcpy.AddMessage(str(fc))
                    desc = arcpy.Describe(fc)
                    state_field_name = "stateName"  # Update with the actual field name for state attribute
                    if state_field_name in desc.fields:
                        #with arcpy.da.SearchCursor(fc, "stateName") as cursor:
                        #with arcpy.da.SearchCursor(fc, [state_field_name]) as cursor:
                        with arcpy.da.SearchCursor(fc, '*') as cursor:
                            field_names = cursor.fields
                            state_field_index = field_names.index(state_field_name)
                            for row in cursor:
                                state = row[state_field_index]
                                arcpy.AddMessage(str(state))
                                arcpy.AddMessage(state)
                                if state and state != "":
                                    if state in state_counts:
                                        state_counts[state] += 1
                                    else:
                                        state_counts[state] = 1
                                else:
                                    arcpy.AddMessage(f"Empty or None value encountered in {state_field_name} field.")
                        # Delete the cursor and row objects
                        del cursor
                        del row
            except Exception as e:
                arcpy.AddMessage(f"An error occurred: {str(e)}")                           
        else:
            arcpy.AddMessage("no if feature_classes")
    except Exception as e:
        arcpy.AddMessage(f"An error occurred while retrieving feature classes: {str(e)}") 
        # fails here after the first feature dataset is scanned

# Define the output CSV file path
output_csv = r"C:\path\output.csv"

# Write the state counts to the CSV file
with open(output_csv, 'w', newline='') as csvfile:
    writer = csv.writer(csvfile)
    writer.writerow(['State', 'Count'])
    for state, count in state_counts.items():
        writer.writerow([state, count])

#arcpy.AddMessage(f"State counts saved to {output_csv}")

if len(state_counts) > 0:
    arcpy.AddMessage(f"State counts saved to {output_csv}")
else:
    arcpy.AddMessage("No valid state values found in the feature classes.")

cleaned up text from View Details window from running as a geoprocessing tool:

workspace: C:\path\data.gdb
[list of feature datasets]
FDS1
Processing feature dataset: FDS1
Number of feature classes: 8
[list of feature classes in FDS]
feature_classes1
2
3...
8

FDS2

Processing feature dataset: FDS2
An error occurred while retrieving feature classes: object of type 'NoneType' has no len()
FDS3
Processing feature dataset: FDS3
An error occurred while retrieving feature classes: object of type 'NoneType' has no len()
CONITNUES HERE OVER EACH FDS...
DOES NOT PRINT ANY MESSAGES ABOUT THE CSV

SeanT · Answer

The script seemed to run through its course, but it got hung up at the end and remained "updating" without finishing properly. Also looking at the output it is not counting as desired.

I don't technically need this anymore for my immediate task, but it would be good to have finished in case it comes up again... with that said, what I am trying to get the tool to count is how many FCs each state is represented in (this is RP asset data), so the output would be each state and like new mexico was found in 80 FCs (out of 172 total FCs in the fgdb). The results in the output were well into the thousands and beyond, so I'm guessing the tool counted features.

In the second version I had posted there was a set created that would increment once for each time the unique state name was found in a FC, not increment once for every time a state was found in the FC. I'm sure a set is one of many ways to do this, and I have no idea/opinion on what is the best way.

Thanks again, great practice for me so far even if I don't manage to finish the tool.

Anonymous User · Answer

Forgot the positional argument for the ListFields() method.  It should be the featureclass so:

if state_field_name in [f.name for f in arcpy.ListFields(fc)]:

Cursors only return the fields that you tell it to and since you do not need / and are not using any other fields than 'stateName', I only passed it to the cursor and it will always be at row[0].

state_field_name = "stateName"
if state_field_name in [f.name for f in arcpy.ListFields(fc)]: # checks if stateName is in the fields and then executes the cursor if its true.
    with arcpy.da.SearchCursor(fc, [state_field_name]) as cursor:

Check out the documentation for the arcpy methods too. It will help you know what is wrong in ChatGPT generated stuff.

SeanT · Answer

I grabbed your script and changed the output csv portion but ran it unchanged otherwise. It now scans each FDS as desired, finds all the FCs in each FDS, but it now raises the following exception for every FDS:

An error occurred while retrieving feature classes: ListFields() missing 1 required positional argument: 'dataset'

The tool ran successfully, but there was no output written to the csv due to each FDS having the exception.

I know that each FDS has FCs, and no FCs are not in a FDS, so those conditions aren't a concern, and the normal organizational fgdb used in the tool would not deviate from that either, fwiw.

In my interactions with chatGPT yesterday, it had recommended scanning the fields for the desired field name of 'stateName' because the first version had row[0]. I know the desired column is not in the exact same index in each FC within the fgdb. Would that matter in what you cleaned up?

Also, thanks so much for the assistance!