How to handle response "None" from GeoPy client using Pandas apply?

Question

I am working on pandas dataframe with quite a few hundreds addresses, trying to add a new column with coordinates received from geopy.

Main question: how to handle unresolved geopy addresses, which gives result "None"?

I am quite new in python and not sure, how to move forward.

My code works, but stops once Latitude/Longitude is not in the database and I get response "None".

Original line:

new_df["coords"] = (
    new_df["address"]
    .progress_apply(geolocator)
    .apply(lambda x: (x.latitude, x.longitude))
)

trying to workout something like the below:

new_df["coords"] = (
    new_df["address"]
    .progress_apply(geolocator)
    .apply(lambda x: np.nan if x == "" else (x.latitude, x.longitude))
)

But I keep receiving error: AttributeError: 'NoneType' object has no attribute 'latitude' and I cannot think of idea how to get around it...

I am testing on 2 addresses at the moment:

1. "Angyalföld - Béke-Tatai utcai lakótelep" - gives result None
2. "Budapest, Bercsényi utca, Hungary" - works properly

full code for testing below, - it will work nicely without the first address:

from random import randint

import pandas as pd
from geopy.exc import *
from geopy.extra.rate_limiter import RateLimiter
from geopy.geocoders import Nominatim
from tqdm import tqdm

tqdm.pandas()  # progress bar
data = ["Budapest, Bercsényi utca, Hungary", "Angyalföld - Béke-Tatai utcai lakótelep"]
df = pd.DataFrame(data, columns=["address"])

user_agent = "geopy_user_{}".format(randint(10000, 99999))
app = Nominatim(user_agent=user_agent)
geolocator = RateLimiter(app.geocode, min_delay_seconds=1)

try:
    df["coords"] = (
        df["address"]
        .progress_apply(geolocator)
        .apply(
            lambda x: (x.latitude, x.longitude)
            if hasattr(x, "latitude") and hasattr(x, "longitude")
            else pd.NA
        )
    )
    print(df)
except GeocoderServiceError as e:
    print("Failed")
    print(e)  # not yet sure how to handle errors - please ignore or advise

Answer 1

One workaround is to prevent apply from failing by checking if x has the attributes "latitude" and "longitude" using Python built-in function hasattr, like this:

df["coords"] = (
    df["address"]
    .progress_apply(geolocator)
    .apply(
        lambda x: (x.latitude, x.longitude)
        if hasattr(x, "latitude") and hasattr(x, "longitude")
        else pd.NA
    )
)