Any method of Replacing NaN values with already existing values

I am working on a dataset in a Jupyter notebook. My dataset has two columns, 'State' and 'City' as you can see in image

Now, State has some missing values in it, So I want to fill NaN data based on the already existing data in the dataset.

For example, in Image 2, the city is Chennai and the state is missing, So I want to fill Tamil Nadu in state column based on Image 1.

Example 2 - I want to fill in the state of Andhra Pradesh, where city is Vishakapatnam and so on for all the NaN data .

As you have now commented "I have one Pandas DF where some state values are missing but available for the same city at other rows in the DF" then you can use groupby on the City and pick-up the state from other row entries for the same City. You can use the code below. The ffill and bfill replaces any NaN values before or after the correct State value.

import pandas as pd


data = {
    'city': ['Delhi', 'Mumbai', 'Bangalore', 'Hyderabad', 'Chennai', 'Delhi', 'Chennai'],
    'state': ['Delhi', 'Maharashtra', 'Karnataka', 'Telangana', None, None, 'Tamil Nadu'],
}

df = pd.DataFrame(data)

df['state'] = df.groupby(['city'])['state'].transform(lambda x: x.ffill().bfill())


print(df)

gives:

        city        state
0      Delhi        Delhi
1     Mumbai  Maharashtra
2  Bangalore    Karnataka
3  Hyderabad    Telangana
4    Chennai   Tamil Nadu
5      Delhi        Delhi
6    Chennai   Tamil Nadu

I believe you need a dictionary that maps each city with its state then it will be a piece of cake.
for India, this link has a JSON file format for all cities and states in India.
Here is a sample code:

data = {
    'city': ['Delhi', 'Mumbai', 'Bangalore', 'Hyderabad', 'Chennai'],
    'state': ['Delhi', 'Maharashtra', 'Karnataka', 'Telangana', 'Tamil Nadu'],
}

city_state_dict = {
    'Delhi': 'Delhi',
    'Mumbai': 'Maharashtra',
    'Bangalore': 'Karnataka',
    'Hyderabad': 'Telangana',
    'Chennai': 'Tamil Nadu',
    'Kolkata': 'West Bengal',
    'Pune': 'Maharashtra',
}

# This line is to simulate the Nan valu in state column
added_row = pd.Series(['Kolkata', np.nan], index=['city', 'state']).to_frame().T
added_row_2 = pd.Series(['Kolkata', 'West Bengal'], index=['city', 'state']).to_frame().T
added_row_3 = pd.Series(['Pune', np.nan], index=['city', 'state']).to_frame().T
final = (
    pd.concat(
        [
            pd.DataFrame(data),
               added_row, added_row_2, added_row_3],
               axis=0,
               ignore_index=True)
    .assign(state=lambda df: df.city.map(city_state_dict))
)
final.head(10)

If you want a quick fix that doesn't cover all cases, you can use forward-fill and back-fill, for instance:

final = (
    pd.concat(
        [
            pd.DataFrame(tt),
               added_row,
               added_row_2,
               added_row_3],
               axis=0,
               ignore_index=True)
    # .assign(state=lambda df: df.city.map(city_state_dict))
    .assign(state= lambda df_: df_.groupby('city', group_key=False).state.apply(lambda x: x.bfill().ffill()))
)
final.head(10)

This will give this output, if there is one city that doesn't have a corresponding state in your data frame but you want to fill it, it will be a NaN with the bfill() and ffill():

Why this happens? because we grouped by city and the city Pune has only one row in the group with no back or forward knowledge about the state to impute it.

I hope this helps!

Example

plz provide code not image and provide desired output

import numpy as np
import pandas as pd

data1 = {'city': ['Delhi', 'Calcutta', 'Delhi', 'Chennai', 'Bokaro', 'Visakhapatnam', 'Chennai', 'Wanparti', 'Delhi', 'Mumbai'], 
         'state': ['Delhi', 'West Bengal', 'Delhi', 'TamilNadu', 'Jharkhand', 'AndhraPradesh', 'TamilNadu', 'AndhraPradesh', 'Delhi', 'Maharashtra']}
df1 = pd.DataFrame(data1)

data2 = {'city': ['Pune', 'Visakhapatnam', 'Indore', 'Visakhapatnam', 'Chennai', 'Chennai', 'Indore', 'Chennai', 'Delhi', 'Pune']}
df2 = pd.DataFrame(data2)
df2['state'] = np.nan

df1

    city	        state
0	Delhi	        Delhi
1	Calcutta	    West Bengal
2	Delhi	        Delhi
3	Chennai	        TamilNadu
4	Bokaro	        Jharkhand
5	Visakhapatnam	AndhraPradesh
6	Chennai	        TamilNadu
7	Wanparti	    AndhraPradesh
8	Delhi	        Delhi
9	Mumbai	        Maharashtra

df2

    city	        state
0	Pune	        NaN
1	Visakhapatnam	NaN
2	Indore	        NaN
3	Visakhapatnam	NaN
4	Chennai	        NaN
5	Chennai	        NaN
6	Indore	        NaN
7	Chennai	        NaN
8	Delhi	        NaN
9	Pune	        NaN

Code

m = dict(df1.values)
df2.assign(state=df2['state'].fillna(df2['city'].map(m)))

output:

    city	        state
0	Pune	        NaN
1	Visakhapatnam	AndhraPradesh
2	Indore	        NaN
3	Visakhapatnam	AndhraPradesh
4	Chennai	        TamilNadu
5	Chennai	        TamilNadu
6	Indore	        NaN
7	Chennai	        TamilNadu
8	Delhi	        Delhi
9	Pune	        NaN