How to Use sort_index in pandas for Sorting DataFrames

Use sort_index() in pandas to sort a DataFrame or Series by its index labels. You can sort in ascending or descending order and choose to sort rows or columns by setting parameters like axis and ascending.

Syntax

The sort_index() method sorts a pandas DataFrame or Series by its index labels.

• axis: Choose 0 or 'index' to sort rows, 1 or 'columns' to sort columns.
• ascending: Set True for ascending order (default), False for descending.
• inplace: If True, modifies the object directly without returning a new one.
• kind: Sorting algorithm, usually left as default.
• na_position: Position of missing values, 'last' or 'first'.

python

DataFrame.sort_index(axis=0, ascending=True, inplace=False, kind='mergesort', na_position='last')

Example

This example shows how to sort a DataFrame by its row index in ascending and descending order, and how to sort columns by their labels.

python

import pandas as pd

# Create a sample DataFrame with unordered index and columns
data = {'B': [4, 3, 2], 'A': [1, 5, 6]}
index = ['c', 'a', 'b']
df = pd.DataFrame(data, index=index)

# Sort rows by index ascending
sorted_rows = df.sort_index()

# Sort rows by index descending
sorted_rows_desc = df.sort_index(ascending=False)

# Sort columns by column labels
sorted_columns = df.sort_index(axis=1)

print('Original DataFrame:\n', df)
print('\nSorted rows ascending:\n', sorted_rows)
print('\nSorted rows descending:\n', sorted_rows_desc)
print('\nSorted columns ascending:\n', sorted_columns)

Output

Original DataFrame:
    B  A
c  4  1
a  3  5
b  2  6

Sorted rows ascending:
    B  A
a  3  5
b  2  6
c  4  1

Sorted rows descending:
    B  A
c  4  1
b  2  6
a  3  5

Sorted columns ascending:
    A  B
c  1  4
a  5  3
b  6  2

Common Pitfalls

One common mistake is forgetting that sort_index() returns a new sorted object by default and does not change the original DataFrame unless inplace=True is set.

Another is mixing up sorting rows (axis=0) and columns (axis=1).

python

import pandas as pd

df = pd.DataFrame({'A': [2, 1]}, index=['b', 'a'])

# Wrong: expecting original df to be sorted without assignment or inplace
sorted_df = df.sort_index()
print('Original df after sort_index without inplace:')
print(df)

# Right: assign back or use inplace
# Option 1: assign
sorted_df = df.sort_index()
print('\nSorted df assigned:')
print(sorted_df)

# Option 2: inplace

df.sort_index(inplace=True)
print('\nOriginal df after inplace sort:')
print(df)

Output

Original df after sort_index without inplace:
   A
b  2
a  1

Sorted df assigned:
   A
a  1
b  2

Original df after inplace sort:
   A
a  1
b  2

Quick Reference

Here is a quick summary of key sort_index() parameters:

Parameter	Description	Default
axis	0 or 'index' to sort rows, 1 or 'columns' to sort columns	0
ascending	Sort order: True for ascending, False for descending	True
inplace	Modify original object if True, else return new sorted object	False
na_position	Position of NaNs: 'last' or 'first'	'last'
kind	Sorting algorithm: 'quicksort', 'mergesort', 'heapsort'	'mergesort'

Key Takeaways

Use sort_index() to sort DataFrame or Series by index labels easily.

Remember sort_index() returns a new object unless inplace=True is set.

Set axis=0 to sort rows and axis=1 to sort columns by their labels.

Use ascending=False to sort in descending order.

Handle missing values position with na_position parameter.