Apply function Pandas ?? Implementation

Question

I made some function like this: data['Age'] = data[['Age',School]].apply(age_implementation, axis = 1) by doing so I want to full NaN values in "Age" column based on School of child and this is the definitione of my "age_implementation" function. Nevertheless when I try to apply function using code above nothing changes, how can I apply this function ?

Function:

def age_implementation(cols):
    Age = cols[0]
    School= cols[1]

    if pd.isnull(Age):
        if School== 1:
            return 10
        elif School== 2:
            return 15
        elif School== 3:
            return 20
        else:
            return Age

Data:

data = pd.DataFrame("School":{1,2,3,1,2,2}, "Age":{NaN, NaN, 20, NaN, NaN, 15})

1. We don't have your data 2. we don't have your code (`age_implementation`) 3. we can't see the output or error messages. Can you please fix these issues so someone can answer your question? — cs95, Dec 13 '19 at 19:20
Yeah, you are assigning Age and School incorrectly. Hint: What is the type of `cols` inside the function? — cs95, Dec 13 '19 at 19:42

score 1 · Answer 1 · answered Dec 13 '19 at 19:35

With how your doing it, i would implement it this way:

def age_implementation(cols):
    Age = cols.Age
    School= cols.School

    if pd.isnull(Age):
        if School== 1:
            return 10
        elif School== 2:
            return 15
        elif School== 3:
            return 20
    else:
         return Age

data['Age'] = data.apply(age_implementation, axis = 1).astype(int)

output:

   School  Age
0       1   10
1       2   15
2       3   20
3       1   10
4       2   15
5       2   15

I think keeping apply is a mistake – ansev Dec 14 '19 at 17:14 — ansev, Dec 14 '19 at 17:14

score 0 · Answer 2 · answered Dec 13 '19 at 19:33

0

We can use Series.map, this is faster:

data['Age']=data['Age'].fillna(data['School'].map({1:10,2:15,3:20}))
print(data)

Output

   School   Age
0       1  10.0
1       2  15.0
2       3  20.0
3       1  10.0
4       2  15.0
5       2  15.0

answered Dec 13 '19 at 19:33

ansev

30,322
5
17
31

great it works! howevew why apply does not work in this case ? – dingaro Dec 13 '19 at 19:35
For their question it seems like `df['School'].add(1).mul(5)` would also work. – cs95 Dec 13 '19 at 19:44
You need : `Age = cols[1] School= cols[0]`. but really if the function is that simple it is not necessary to use `apply`. if the logic is linear then you can use what it proposes @cs95 – ansev Dec 13 '19 at 19:46
Series.map is faster – ansev Dec 13 '19 at 19:46

Apply function Pandas ?? Implementation

2 Answers2