Read Values other than NaN as NaN

Not all null values have NaN in them some may have other substitutes like ‘_’ or ‘?’ To mitigate that we have na_values in read_csv

import pandas as pd
d = {'id':[0, 1, 2], 'Name':['Ben', 'Max', '?']}
df = pd.DataFrame(d)
df.head()
id Name
0 0 Ben
1 1 Max
2 2 ? 
df.isnull().sum().any()
False

’?’ is interpreted as not a null value here

df.to_csv('test.csv')
df_real = pd.read_csv('test.csv', na_values=['?']) # Reads '?' as NaN value
df_real.head()
Unnamed: 0 id Name
0 0 0 Ben
1 1 1 Max
2 2 2 NaN 
df_real.isnull().sum().any()
True