Add missing data, Replace missing data , transform missing data, remove rows with missing data using fillna() in pandas library or by using Imputer in sklear.preprocessing.Imputer

Filling,Removing or processing missing data from tabular dataset

import pandas as pd 
import numpy as np 

df=pd.read_csv("/home/pima_indians_diabetes.csv",header=None)

# # to get a statistical summary of all columns, so that we can identify missing data by checking min value
# print(df.describe())

# print(df.head(20))

# # to find number of missing data in each column 
# print((df[[0, 1, 2, 3, 4, 5, 6, 7, 8]]==0).sum())

# # to replace missing values with NaN 
df[[0, 1, 2, 3, 4, 5, 6, 7, 8]] = df[[0, 1, 2, 3, 4, 5, 6, 7, 8]].replace(0,np.NaN)


# # count number of NaN values in each column to check if it is replaced
# print(df[[0, 1, 2, 3, 4, 5, 6, 7, 8]].isnull().sum())
# print(df.head(20))

# #  drop rows with NaN values using dropna()
# print("Original dataset shape is {}".format(df.shape))

# # df.dropna(inplace=True)
# print("Reduced dataset shape after removing rows with NaN is {}".format(df.shape))

# # using pandas fillna() to impute or add missing values
# df.fillna(df.mean(), inplace=True)
# print(df.head(20))

# # count the number of NaN values in each column
# print(df.isnull().sum())

# using imputer in sklearn.preprocessing to add missing values
from sklearn.preprocessing import Imputer

imputer=Imputer(missing_values = 'NaN', strategy = 'mean', axis = 0)
values=df.values

transformrd=imputer.fit_transform(values)
print(transformrd)