# Having the following Dataframe:

   First  Second  Third  Fourth
0    NaN       2    NaN       0
1      3       4    NaN       1
2    NaN     NaN    NaN       5
3      4       1      2       3


# by default dropna() removes rows with NaN in ANY of the colums

df.dropna()

   First  Second  Third  Fourth
3      4       1      2       3

# Remove NaN only from an specific column

df.dropna(subset=['Second'])

   First  Second  Third  Fourth
0    NaN       2    NaN       0
1      3       4    NaN       1
3      4       1      2       3