How to use phonenumbers Python library on a pandas dataframe testing phone field before?-CodePudding

I would to parse a phone field in a pandas dataframe. My phone field is named Phone and my country code field CountryCode.

It works but only when all the phones are filled:

df['Test phone'] = df.apply(lambda x:phonenumbers.parse(str(x['Phone']), str(x['CountryCode'])), axis='columns')

OK, but how to do it only if the phone, or even the country code, are filled? I tried a lot of if/else syntaxes in the lambda function, without success.

Amazing, I also tried doing it in a df.loc, it does not work neither...

Please, have you an idea?

THX

CodePudding user response：

Just use an if and else statement within your lambda and test with row.notnull().all() if all row entries are not null.

I added an example from the phonenumbers project.


import numpy as np
import pandas as pd
import phonenumbers

df = pd.DataFrame(
    [
        [
            np.nan,
            "020 8366 1177",
        ],
        ["GB", "020 8366 1177"],
        [np.nan, np.nan],
    ],
    columns=["CountryCode", "Phone"],
)

df["Test phone"] = df.apply(
    lambda row: phonenumbers.parse(str(row["Phone"]), str(row["CountryCode"]))
    if row.notnull().all()
    else "",
    axis="columns",
)

Output:

    CountryCode Phone   Test phone
0   NaN 020 8366 1177   
1   GB  020 8366 1177   Country Code: 44 National Number: 2083661177
2   NaN NaN

CodePudding user response：

Thank you

Your code do not work as it, but you very helped me: it was a syntax error.

From your code, now I can do:

df['Phone'] = df['Phone'].fillna('')
df['CountryCode'] = df['CountryCode'].fillna('')

df['TestPhone'] = df.apply(
    lambda x: phonenumbers.parse(str(x['Phone']), str(x['CountryCode']))
    if str(x['Phone']) != ''
    and str(x['CountryCode']) != ''
    else '',
    axis='columns'
)