stack() and unstack() in Pandas - Step-by-Step Execution

Concept Flow - stack() and unstack()

Start with DataFrame

↓

Call stack()

↓

Stacked Series with MultiIndex

↓

Call unstack()

↓

Original or reshaped DataFrame

stack() turns columns into rows creating a MultiIndex Series; unstack() reverses this by turning row levels back into columns.

Execution Sample

Pandas

import pandas as pd

df = pd.DataFrame({
  'A': [1, 2],
  'B': [3, 4]
})
stacked = df.stack()
unstacked = stacked.unstack()

This code creates a DataFrame, stacks it to turn columns into rows, then unstacks it back to the original shape.

Execution Table

Step	Action	Data Structure	Shape/Index	Output Preview
1	Create DataFrame df	DataFrame	(2 rows, 2 columns)	A B 0 1 3 1 2 4
2	Call df.stack()	Series with MultiIndex	(4 rows)	0 A 1 B 3 1 A 2 B 4
3	Call stacked.unstack()	DataFrame	(2 rows, 2 columns)	A B 0 1 3 1 2 4
4	End	-	-	stack() and unstack() complete, DataFrame restored

💡 All steps executed; stack() converted columns to rows, unstack() reversed it.

Variable Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	Final
df	None	DataFrame(2x2)	DataFrame(2x2)	DataFrame(2x2)	DataFrame(2x2)
stacked	None	None	Series with MultiIndex(4)	Series with MultiIndex(4)	Series with MultiIndex(4)
unstacked	None	None	None	DataFrame(2x2)	DataFrame(2x2)

Key Moments - 3 Insights

Why does stack() result in a Series with a MultiIndex instead of a DataFrame?

stack() compresses columns into a single row index level, creating a Series with a MultiIndex to keep track of original row and column labels, as shown in execution_table step 2.

What happens if you call unstack() on a Series without a MultiIndex?

unstack() requires a MultiIndex to pivot one level back to columns. Without it, unstack() will raise an error or not reshape properly, unlike the successful unstack in step 3.

Does unstack() always restore the original DataFrame shape?

If no data is lost or filtered, unstack() reverses stack() perfectly, restoring the original shape as in step 3. But if data is missing or filtered, the shape may differ.

Visual Quiz - 3 Questions

Look at the execution table at step 2. What is the shape of the stacked data?

A 4 rows, single column Series with MultiIndex

B 2 rows, 2 columns DataFrame

C 4 rows, 2 columns DataFrame

D 2 rows, single column Series

💡 Hint

Refer to the 'Shape/Index' and 'Data Structure' columns in step 2 of the execution table.

At which step does the data return to its original DataFrame shape?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Check the 'Shape/Index' and 'Output Preview' columns for step 3 in the execution table.

If the original DataFrame had 3 columns instead of 2, how would the stacked Series length change at step 2?

A It would have 3 rows

B It would have 6 rows

C It would remain 4 rows

D It would have 9 rows

💡 Hint

Stacking multiplies rows by columns; see variable_tracker for how shape changes.

Concept Snapshot

stack(): Converts DataFrame columns into rows, creating a Series with MultiIndex.
unstack(): Reverses stack(), turning MultiIndex rows back into columns.
Use stack() to reshape wide data to long format.
Use unstack() to reshape long data back to wide.
Both preserve data labels via MultiIndex.
Common in data cleaning and reshaping tasks.

Full Transcript

This visual execution shows how pandas stack() and unstack() work. Starting with a DataFrame of 2 rows and 2 columns, stack() compresses columns into rows, creating a Series with a MultiIndex that tracks original row and column labels. Then unstack() reverses this operation, restoring the original DataFrame shape. The execution table traces each step with data structure type, shape, and output preview. Variable tracker shows how variables df, stacked, and unstacked change. Key moments clarify common confusions about MultiIndex and shape restoration. The quiz tests understanding of shapes and transformations. This helps beginners see exactly how stack and unstack reshape data step-by-step.