When I need to apply successive steps of data transformation to a DataFrame (i.e. basically in every data analysis), I favor method chaining for better readability and efficiency. However, some operations stayed rather obscure to me for a long time, like renaming all columns in a single operation. Here are some tips that may come in handy.

Setup

# Import library
import pandas as pd

# Create sample DataFrame
df = pd.DataFrame({'First Name': ['John', 'Aby', 'Bob', 'Alice'],
                  'Last Name': ['Doe', 'Parker', 'Morris', 'Allen']})

df

	First Name	Last Name
0	John	Doe
1	Aby	Parker
2	Bob	Morris
3	Alice	Allen

Classic, non-chained method

A standard way of applying transformative functions to all columns names would be the following:

# Classic way
df2 = df.copy()
df2.columns = [col.lower().replace(' ', '_') for col in df2.columns]
df2

	first_name	last_name
0	John	Doe
1	Aby	Parker
2	Bob	Morris
3	Alice	Allen

Apply transformative functions

With pandas method chaining, you can apply functions to transform names with .rename(columns=<function>).

Applying a function that does not require any argument, like lower() or title(), is very short:

# Method chaining: convert to lowercase
df.rename(columns=str.lower)

	first name	last name
0	John	Doe
1	Aby	Parker
2	Bob	Morris
3	Alice	Allen

For a function that requires arguments, like str.replace(), use a lambda:

# Method chaining: replace strings
df.rename(columns=lambda x: x.replace(' ', ''))

	FirstName	LastName
0	John	Doe
1	Aby	Parker
2	Bob	Morris
3	Alice	Allen

Specify new names to columns

If you want to specify a list of new columns names, use set_axis():

# Method chaining: rename all columns
df.set_axis(['given_name', 'family_name'], axis=1)

	given_name	family_name
0	John	Doe
1	Aby	Parker
2	Bob	Morris
3	Alice	Allen

Rename pandas DataFrame columns with method chaining

Setup

Classic, non-chained method

Apply transformative functions

Specify new names to columns