Panda: Using fillna() with specific columns in a DataFrame

# Table of Contents

# Panda: Using fillna() with a specific column in a DataFrame

To use the fillna() method with only specific columns in a Pandas DataFrame:

Select the DataFrame column(s) using bracket notation.
Call the fillna() method on the selected column(s).

main.py

Copied!
import pandas as pd

df = pd.DataFrame({
    'ID': [1, 1, None, 2, 2, None],
    'Animal': ['Cat', 'Cat', None, 'Dog', 'Dog', None],
    'Max Speed': [25, 35, None, 55, 65, None]
})

print(df)

print('-' * 50)

df['Animal'] = df['Animal'].fillna(value='Anonymous')

print(df)

The code for this article is available on GitHub

Running the code sample produces the following output.

shell

Copied!
   ID Animal  Max Speed
0   1    Cat       25.0
1   1    Cat       35.0
2   1   None        NaN
3   2    Dog       55.0
4   2    Dog       65.0
5   2   None        NaN
--------------------------------------------------
   ID     Animal  Max Speed
0   1        Cat       25.0
1   1        Cat       35.0
2   1  Anonymous        NaN
3   2        Dog       55.0
4   2        Dog       65.0
5   2  Anonymous        NaN

fillna only specific column in dataframe

We used bracket notation [] to select the "Animal" column and used the DataFrame.fillna method to fill the NA/NaN values in the column with the string "Anonymous".

main.py

Copied!
df['Animal'] = df['Animal'].fillna(value='Anonymous')

The value you pass to the fillna() method is the replacement for each NA/NaN value in the specified column.

# Panda: Using fillna() with multiple specific columns in a DataFrame

If you need to use the fillna() method with multiple specific columns, use two sets of square brackets when selecting them.

main.py

Copied!
import pandas as pd

df = pd.DataFrame({
    'ID': [1, 1, None, 2, 2, None],
    'Animal': ['Cat', 'Cat', None, 'Dog', 'Dog', None],
    'Max Speed': [25, 35, None, 55, 65, None]
})

print(df)

print('-' * 50)

df[['Animal', 'Max Speed']] = df[[
    'Animal', 'Max Speed']].fillna(value='NOT_AVAILABLE')

print(df)

The code for this article is available on GitHub

Running the code sample produces the following output.

shell

Copied!
    ID Animal  Max Speed
0  1.0    Cat       25.0
1  1.0    Cat       35.0
2  NaN   None        NaN
3  2.0    Dog       55.0
4  2.0    Dog       65.0
5  NaN   None        NaN
--------------------------------------------------
    ID         Animal      Max Speed
0  1.0            Cat           25.0
1  1.0            Cat           35.0
2  NaN  NOT_AVAILABLE  NOT_AVAILABLE
3  2.0            Dog           55.0
4  2.0            Dog           65.0
5  NaN  NOT_AVAILABLE  NOT_AVAILABLE

using fillna with multiple specific columns in dataframe

Notice that two sets of square brackets [[]] are used when selecting multiple columns.

main.py

Copied!
df[['Animal', 'Max Speed']] = df[[
    'Animal', 'Max Speed']].fillna(value='NOT_AVAILABLE')

Once we've selected the multiple columns, we can call the fillna() method with the replacement.

The code sample used the "NOT_AVAILABLE" string for the replacement, but you can use any other value.

The value doesn't necessarily have to be a string.

You can also use the DataFrame.loc label-based indexer, however, this isn't necessary.

main.py

Copied!
import pandas as pd

df = pd.DataFrame({
    'ID': [1, 1, None, 2, 2, None],
    'Animal': ['Cat', 'Cat', None, 'Dog', 'Dog', None],
    'Max Speed': [25, 35, None, 55, 65, None]
})

print(df)

print('-' * 50)

df[['Animal', 'Max Speed']] = df.loc[:, [
    'Animal', 'Max Speed']].fillna(value='NOT_AVAILABLE')

print(df)

The code for this article is available on GitHub

Running the code sample produces the same output.

shell

Copied!
    ID Animal  Max Speed
0  1.0    Cat       25.0
1  1.0    Cat       35.0
2  NaN   None        NaN
3  2.0    Dog       55.0
4  2.0    Dog       65.0
5  NaN   None        NaN
--------------------------------------------------
    ID         Animal      Max Speed
0  1.0            Cat           25.0
1  1.0            Cat           35.0
2  NaN  NOT_AVAILABLE  NOT_AVAILABLE
3  2.0            Dog           55.0
4  2.0            Dog           65.0
5  NaN  NOT_AVAILABLE  NOT_AVAILABLE

fillna only specific columns in dataframe using loc

# Panda: Using fillna() with specific columns in a DataFrame using a `dict`

You can also pass a dictionary to the fillna() method to only call the method on specific columns.

main.py

Copied!
import pandas as pd

df = pd.DataFrame({
    'ID': [1, 1, None, 2, 2, None],
    'Animal': ['Cat', 'Cat', None, 'Dog', 'Dog', None],
    'Max Speed': [25, 35, None, 55, 65, None]
})

print(df)

print('-' * 50)

df = df.fillna(
    {'Animal': 'Anonymous',
     'Max Speed': 'NOT_AVAILABLE'}
)

print(df)

The code for this article is available on GitHub

Running the code sample produces the following output.

shell

Copied!
    ID Animal  Max Speed
0  1.0    Cat       25.0
1  1.0    Cat       35.0
2  NaN   None        NaN
3  2.0    Dog       55.0
4  2.0    Dog       65.0
5  NaN   None        NaN
--------------------------------------------------
    ID     Animal      Max Speed
0  1.0        Cat           25.0
1  1.0        Cat           35.0
2  NaN  Anonymous  NOT_AVAILABLE
3  2.0        Dog           55.0
4  2.0        Dog           65.0
5  NaN  Anonymous  NOT_AVAILABLE

fillna specific columns using dict

We used the column names as dictionary keys and the replacements as dictionary values.

main.py

Copied!
df = df.fillna(
    {'Animal': 'Anonymous',
     'Max Speed': 'NOT_AVAILABLE'}
)

You can include as many key-value pairs in the dictionary as necessary.

The code sample above updates the DataFrame by reassigning the variable.

Alternatively, you can mutate the DataFrame directly by setting the inplace argument to True.

main.py

Copied!
import pandas as pd

df = pd.DataFrame({
    'ID': [1, 1, None, 2, 2, None],
    'Animal': ['Cat', 'Cat', None, 'Dog', 'Dog', None],
    'Max Speed': [25, 35, None, 55, 65, None]
})

print(df)

print('-' * 50)

df.fillna(
    {'Animal': 'Anonymous',
     'Max Speed': 'NOT_AVAILABLE'},
    inplace=True
)

print(df)

The code for this article is available on GitHub

Running the code sample produces the following output.

shell

Copied!
    ID Animal  Max Speed
0  1.0    Cat       25.0
1  1.0    Cat       35.0
2  NaN   None        NaN
3  2.0    Dog       55.0
4  2.0    Dog       65.0
5  NaN   None        NaN
--------------------------------------------------
    ID     Animal      Max Speed
0  1.0        Cat           25.0
1  1.0        Cat           35.0
2  NaN  Anonymous  NOT_AVAILABLE
3  2.0        Dog           55.0
4  2.0        Dog           65.0
5  NaN  Anonymous  NOT_AVAILABLE

fillna specific columns in place

When the inplace argument is set to True, the fillna() method mutates the DataFrame in place and returns None.

main.py

Copied!
df.fillna(
    {'Animal': 'Anonymous',
     'Max Speed': 'NOT_AVAILABLE'},
    inplace=True
)

# Additional Resources

You can learn more about the related topics by checking out the following tutorials:

I wrote a book in which I share everything I know about how to become a better, more efficient programmer.

You can use the search field on my Home Page to filter through all of my articles.

Panda: Using fillna() with specific columns in a DataFrame

# Table of Contents

# Panda: Using fillna() with a specific column in a DataFrame

# Panda: Using fillna() with multiple specific columns in a DataFrame

# Panda: Using fillna() with specific columns in a DataFrame using a dict

# Additional Resources

Borislav Hadzhiev

# Panda: Using fillna() with specific columns in a DataFrame using a `dict`