missing value conditions Pandas in a function

Question

I would like a function where if the area column has missing values (like NULL in SQL) the result is 'A' in the target 'wanted' variable.

I'm confused about use of None, isnull(), np.nan concepts in Python


raw_data = {'area': ['S','W',np.nan,np.nan], 'wanted': [np.nan,np.nan,'A','A']}
df = pd.DataFrame(raw_data, columns = ['area','wanted'])
df


def my_func(x):
    if (x) is None:
        return 'A'
    else:
        return np.nan


df['wanted2'] = df['area'].apply(my_func)

df

score 3 · Accepted Answer · answered Apr 05 '20 at 13:38

3

np.nan is not equal to None , infact NaN isnot equal to NaN as well (check np.nan == None) , hence you can utilize pd.isna() in your if condition:

def my_func(x):
    if pd.isna(x):
        return 'A'
    else:
        return np.nan


df['wanted2'] = df['area'].apply(my_func)

but you can vectorize this using np.where and series.isna() instead of using apply

df['wanted2'] = np.where(df['area'].isna(),'A',np.nan)

answered Apr 05 '20 at 13:38

anky

74,114
11
41
70

thanks and just to know, the opposite of "pd.isna(x)", not having NULL, how is it? – progster Apr 05 '20 at 13:50
@progster `pd.notna(x)` , similarly for a series , we have `series.notna()` – anky Apr 05 '20 at 13:52

score 0 · Answer 2 · answered Apr 05 '20 at 13:29

0

You can use fill.na

df['wanted2'] = df.area.fillna('A')

In your code return np.nan if the value exists in area and 'A' otherwise.

answered Apr 05 '20 at 13:29

Tom Ron

5,906
3
22
38

thanks but in order to understand I would like to do it wih my function, because this is only an example, in real life the conditions are more complex – progster Apr 05 '20 at 13:35
Given your input, was is your expect output? – Tom Ron Apr 05 '20 at 13:37
my target was having something like the 'wanted' variable – progster Apr 05 '20 at 13:45

missing value conditions Pandas in a function

2 Answers2