Swapping index levels in Pandas - Step-by-Step Execution

Concept Flow - Swapping index levels

Start with MultiIndex DataFrame

↓

Choose levels to swap

↓

Call swaplevel(level1, level2)

↓

New DataFrame with swapped index levels

↓

Use swapped DataFrame for further operations

Swapping index levels means exchanging positions of two levels in a MultiIndex DataFrame to change how data is grouped or accessed.

Execution Sample

Pandas

import pandas as pd
idx = pd.MultiIndex.from_tuples([('A', 1), ('A', 2), ('B', 1)])
df = pd.DataFrame({'val': [10, 20, 30]}, index=idx)
df_swapped = df.swaplevel(0, 1)

This code creates a DataFrame with a two-level index and swaps the two index levels.

Execution Table

Step	Action	Index Before	Index After	Resulting DataFrame Index
1	Create MultiIndex from tuples [('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]	MultiIndex with level0: ['A','A','B'], level1: [1,2,1]
2	Create DataFrame with MultiIndex	[('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]	DataFrame with index levels 0:'A','A','B' and 1:1,2,1
3	Call swaplevel(0,1) to swap index levels	[('A',1), ('A',2), ('B',1)]	[(1,'A'), (2,'A'), (1,'B')]	DataFrame index levels swapped: level0: [1,2,1], level1: ['A','A','B']
4	Use swapped DataFrame for further operations	[(1,'A'), (2,'A'), (1,'B')]	[(1,'A'), (2,'A'), (1,'B')]	DataFrame now grouped/indexed by swapped levels

💡 All steps complete; index levels successfully swapped

Variable Tracker

Variable	Start	After Step 1	After Step 3	Final
idx	None	[('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]
df.index	None	None	[('A',1), ('A',2), ('B',1)]	[('A',1), ('A',2), ('B',1)]
df_swapped.index	None	None	[(1,'A'), (2,'A'), (1,'B')]	[(1,'A'), (2,'A'), (1,'B')]

Key Moments - 2 Insights

Why does the order of index levels matter after swapping?

The order of index levels affects how data is grouped and accessed. After swapping (see execution_table step 3), the first level becomes the previous second level, changing grouping and sorting behavior.

Does swaplevel change the data or just the index structure?

swaplevel only changes the order of index levels, not the data itself. The rows remain the same but are accessed differently (see execution_table step 3 and 4).

Visual Quiz - 3 Questions

Look at the execution_table at step 3, what is the new first index level after swaplevel(0,1)?

A A combination of both levels

B The original first level values ['A', 'A', 'B']

C The original second level values [1, 2, 1]

D An empty index

💡 Hint

Check the 'Index After' column at step 3 in execution_table

At which step does the DataFrame index actually change to swapped levels?

A Step 3

B Step 2

C Step 1

D Step 4

💡 Hint

Look at the 'Resulting DataFrame Index' column in execution_table

If you swap levels back again after step 3, what will the index look like?

A Swapped index [(1,'A'), (2,'A'), (1,'B')]

B Same as original index [('A',1), ('A',2), ('B',1)]

C Index with only one level

D Index with levels reversed but values changed

💡 Hint

Swapping twice returns index to original order (see variable_tracker for df_swapped.index)

Concept Snapshot

swaplevel(level1, level2) swaps two index levels in a MultiIndex DataFrame.
It changes the order of index levels but not the data.
Useful for changing grouping or sorting order.
Returns a new DataFrame with swapped index levels.
Example: df.swaplevel(0,1) swaps first and second index levels.

Full Transcript

This visual execution shows how to swap index levels in a pandas MultiIndex DataFrame. We start with a MultiIndex created from tuples [('A',1), ('A',2), ('B',1)]. The DataFrame uses this MultiIndex as its index. When we call swaplevel(0,1), the first and second index levels exchange places. The index changes from level0: ['A','A','B'] and level1: [1,2,1] to level0: [1,2,1] and level1: ['A','A','B']. This affects how the DataFrame is grouped and accessed but does not change the data itself. Swapping again would restore the original index order. This technique helps reorganize data views without altering the underlying data.