Is filled with objects: DataFrame
A number or a character string to fill
df.fillna (0) or df.fillna ( ',')
2 is filled with data prior to a
df.fillna( method = 'pad')
3 is filled with a data
df.fillna(method='bfill',limit=1)
With restrictions limit the number of possible alternative NaN each column, let's limit can only replace a NaN each column
4 using the average or other descriptive statistics replace missing values
df.fillna( df.mean() )
5 select which column of data is processed missing values
df.fillna( df.mean()['one','two'] )