Search code examples
pythondataframelambda

DF fillna a column according to one hot columns mean

I'm working with Titanic data set and wanted to fill age's na values according to tittles that i get using the names. I want to fill that age value which is in the group(miss, mrs, mr)'s own age mean

age Tittle_Mr Tittle_Miss Tittle_Mrs
22 1 0 0
38 0 1 0
26 0 0 1
35 1 0 0
NaN 1 0 0

I want that na to be Tittle_Mr's age's mean.

In this case 57/2 ~ 28

Solution 
  • import pandas as pd
import numpy as np

df = pd.DataFrame({'age': [22 , 38,  26,35,np.nan],"Tittle_Mr":[1,0,0,1,1],"Tittle_Miss":[0,1,0,0,0],"Tittle_Mrs":[0,0,1,0,0]})

indexs=list(np.where(df['age'].isnull())[0])

for i in indexs:
   if df["Tittle_Mr"][i]==1:
        df["age"][i]=np.mean(df.where(df["Tittle_Mr"]==1))["age"]
   elif df["Tittle_Miss"][i]==1:
        df["age"][i]=np.mean(df.where(df["Tittle_Miss"]==1))["age"]
   else:
        df["age"][i]=np.mean(df.where(df["Tittle_Mrs"]==1))["age"]

 print(df)

    please try this