pandas add new column based on conditional logic of missing values

Question

I have a pandas dataframe like this:

aa bb   cc dd ee
a  a    b  b  foo
a  b    a  a  foo
b  nan  a  a  bar
b  b    b  b  bar

I want to create a new column df['ff'] like:

aa bb   cc dd ee   ff
a  a    b  b  foo  c
a  b    a  a  foo  c
a  nan  a  a  bar  d
a  b    b  b  bar  c

The logic is: if df['bb'] is not null and df['aa']==a, then c else d

Based on answers to other questions, I think the answer should be something like this:

df['ff'] = df.apply(lambda x: x['bb'].isnull(),axis=1) & (x['aa']=='a')

But I get an error like this:

("'str' object has no attribute 'isnull'", 'occurred at index 0')

Answer 1

I'd use the following vectorized approach:

In [47]: df['ff'] = np.where(df['bb'].notnull() & df['aa'].eq('a'), 'c', 'd')

In [48]: df
Out[48]:
  aa   bb cc dd   ee ff
0  a    a  b  b  foo  c
1  a    b  a  a  foo  c
2  b  NaN  a  a  bar  d
3  b    b  b  b  bar  d