MultiIndex (hierarchical indexing) in Data Analysis Python - Step-by-Step Execution

Concept Flow - MultiIndex (hierarchical indexing)

Create MultiIndex from tuples

↓

Assign MultiIndex to DataFrame

↓

Access data by outer level

↓

Access data by inner level

↓

Slice or filter using MultiIndex

We build a MultiIndex from tuples, assign it to a DataFrame, then access or slice data by levels.

Execution Sample

Data Analysis Python

import pandas as pd
index = pd.MultiIndex.from_tuples([('A', 1), ('A', 2), ('B', 1)])
df = pd.DataFrame({'Value': [10, 20, 30]}, index=index)
print(df.loc['A'])

Create a DataFrame with a MultiIndex and select rows where the first level is 'A'.

Execution Table

Step	Action	Variable/Expression	Result/Value
1	Create MultiIndex from tuples	pd.MultiIndex.from_tuples([('A',1),('A',2),('B',1)])	MultiIndex([('A', 1), ('A', 2), ('B', 1)], names=[None, None])
2	Create DataFrame with MultiIndex	pd.DataFrame({'Value':[10,20,30]}, index=index)	DataFrame with MultiIndex: Value A 1 10 2 20 B 1 30
3	Access rows with first level 'A'	df.loc['A']	DataFrame: Value 1 10 2 20
4	Access rows with second level 1 under 'A'	df.loc[('A',1)]	Series: Value 10 Name: (A, 1), dtype: int64
5	Slice rows where second level >= 2	df.loc[pd.IndexSlice[:, 2:], :]	DataFrame: Value A 2 20
6	Exit	No more operations	End of example

💡 All steps executed to show MultiIndex creation, assignment, and access.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	Final
index	None	MultiIndex([('A', 1), ('A', 2), ('B', 1)])	MultiIndex([('A', 1), ('A', 2), ('B', 1)])	Same	Same	Same	Same
df	None	None	DataFrame with MultiIndex and values	Same	Same	Same	Same
df.loc['A']	None	None	None	DataFrame with rows for 'A'	Same	Same	Same
df.loc[('A',1)]	None	None	None	None	Series with value 10	Same	Same
df.loc[pd.IndexSlice[:, 2:], :]	None	None	None	None	None	DataFrame with rows where second level >= 2	Same

Key Moments - 3 Insights

Why does df.loc['A'] return multiple rows?

Because 'A' is the first level of the MultiIndex and selecting it returns all rows where the first level is 'A', as shown in execution_table step 3.

What happens if you try df.loc['C'] which is not in the index?

It raises a KeyError because 'C' is not present in the first level of the MultiIndex, similar to how normal indexing works but with hierarchical levels.

How does slicing with pd.IndexSlice[:, 2:] work?

It selects all rows where the second level of the MultiIndex is 2 or greater, as shown in execution_table step 5.

Visual Quiz - 3 Questions

Look at the execution_table step 3, what rows does df.loc['A'] return?

A All rows in the DataFrame

B Rows where second level is 1

C Rows where first level is 'A' (two rows)

D Only the row ('A', 1)

💡 Hint

Check the 'Result/Value' column in step 3 of execution_table.

At which step does df.loc[('A',1)] return a Series?

A Step 2

B Step 4

C Step 5

D Step 3

💡 Hint

Look at the 'Action' and 'Result/Value' columns in execution_table for step 4.

If we change the tuple ('B', 1) to ('B', 3), which step's output changes?

A Step 5 output changes

B Step 3 output changes

C Step 4 output changes

D No output changes

💡 Hint

Step 5 filters rows where second level >= 2, so changing ('B',1) to ('B',3) affects this.

Concept Snapshot

MultiIndex (hierarchical indexing) lets you use multiple levels of row labels.
Create with pd.MultiIndex.from_tuples() and assign as DataFrame index.
Access data by levels using df.loc[level1] or df.loc[(level1, level2)].
Use pd.IndexSlice for slicing across levels.
It helps organize complex data with multiple keys.

Full Transcript

This example shows how to create a MultiIndex from tuples and assign it to a DataFrame. We then access data by selecting the first level 'A', which returns multiple rows. Selecting a full tuple like ('A',1) returns a single row as a Series. We also slice rows where the second level is greater or equal to 2 using pd.IndexSlice. The variable tracker shows how variables change after each step. Key moments clarify common confusions about selecting levels and slicing. The visual quiz tests understanding of these steps. MultiIndex helps organize data with multiple keys for easier access and slicing.