error: 'str' object does not support item assignment in python

Question

I am trying to geocode address using geopy by using some code downloaded from github. but the error the code is that the str obj doesn't support item assignment in geocode_addresses address_dict["error"] = ""? why do I get that error. below is the code.

if __name__ == '__main__':
    csv_file = 'nr-fixedNew.csv'
    with open(csv_file, 'rb') as csvfile:
        for row in csvfile.readlines():
            df = pd.read_csv(csv_file)
            address = df.ADDRESS
            geocoded = geocode_addresses(address)
            write_csv(output_file, geocoded)


    def geocode_addresses(address_dicts):
        geocoder = geocoders.GoogleV3()
        for address_dict in address_dicts:
            address_dict["error"] = ""
            try:
                time.sleep(1)
                address, (lat, lon) = geocoder.geocode(address_dict["fulladdress"])
                address_dict["fulladdress"] = address
                address_dict["latitude"] = lat
                address_dict["longitude"] = lon
            except ValueError as e:
                address_dict["error"] = e
        return address_dicts

dataset example

SEX,PROGRAMME,ADDRESS
M,2,"J6855, JALAN LMBAH KESANG 1/1-3,77378 MERLIMAU, MELAKA";

Answer 1

Here it looks like address/df.ADDRESS in the snippet below:

    address = df.ADDRESS
    geocoded = geocode_addresses(address)

address is not a list of dicts like you're expecting it to be:

for address_dict in address_dicts:
    address_dict["error"] = ""

---

after testing your code, here's what address_dicts is:

(<class 'pandas.core.series.Series'>

which is basically a (pandas) list of strings:

(Pdb) address_dicts
0    J6855, JALAN LMBAH KESANG 1/1-3,77378 MERLIMAU...
Name: ADDRESS, dtype: object
(Pdb) address_dicts[0]
'J6855, JALAN LMBAH KESANG 1/1-3,77378 MERLIMAU, MELAKA;'

so your issue is related to how you're parsing the CSV file.

---

And tbh, I don't really understand what you're trying to do here:

with open(csv_file, 'rb') as csvfile:
    for row in csvfile.readlines():
        df = pd.read_csv(csv_file)

so you're opening the file, then you iterate over the file, and for each line you're parsing the file as a csv ?!

---

if __name__ == '__main__':
    csv_file = 'nr-fixedNew.csv'
    with open(csv_file, 'rb') as csvfile:
            df = pd.read_csv(csv_file).to_dict() ### here you convert to dict
            address = df['ADDRESS']
            geocoded = geocode_addresses(address)
            write_csv(output_file, geocoded)

and then you need to rethink the way you're parsing the values. Don't hesitate to call pdb or add printouts.