You can also specify missing values during DataFrame creation, specially if you are reading this from a file.

import pandas as pd

df = pd.read_csv("some_file.csv", na_values=["?"])