treatment of null value on a dataset-CodePudding

I have a data set where null value is

df.isnull().sum()
country                       0
country_long                  0
name                          0
gppd_idnr                     0
capacity_mw                   0
latitude                     46
longitude                    46
primary_fuel                  0
other_fuel1                   0
other_fuel2                   0
other_fuel3                 908
commissioning_year          380
owner                         0
source                        0
url                           0
geolocation_source            0
wepp_id                     908
year_of_capacity_data       388
generation_gwh_2013         524
generation_gwh_2014         507
generation_gwh_2015         483
generation_gwh_2016         471
generation_gwh_2017         465
generation_data_source        0
estimated_generation_gwh    908

I tried mean mode max min and std all the methods but all null values is not removing when I try

df['wepp_id']=df['wepp_id'].replace(np.NAN,df['wepp_id'].mean())

its not working same things happen on median , std and min, max also

CodePudding user response：

Try df['wepp_id']=df['wepp_id'].fillna(df['wepp_id'].mean())

If this does not work, then it means that your column is not of number type. If it is an string type, then you need to do this first:

df['wepp_id'] = df['wepp_id'].astype(float)

Then run the command in the first line.