Stack and unstack in Data Analysis Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ data_output

intermediate

Output of stacking a DataFrame with multi-level columns

Given the following DataFrame with multi-level columns, what is the output after applying df.stack(level=0).stack()?

Data Analysis Python

import pandas as pd

cols = pd.MultiIndex.from_tuples([('A', 'one'), ('A', 'two'), ('B', 'one'), ('B', 'two')])
data = [[1, 2, 3, 4], [5, 6, 7, 8]]
df = pd.DataFrame(data, columns=cols)
result = df.stack(level=0).stack()

A

0  A  one    1
         two    2
   B  one    3
         two    4
1  A  one    5
         two    6
   B  one    7
         two    8
dtype: int64

B {0: {('A', 'one'): 1, ('A', 'two'): 2, ('B', 'one'): 3, ('B', 'two'): 4}, 1: {('A', 'one'): 5, ('A', 'two'): 6, ('B', 'one'): 7, ('B', 'two'): 8}}

C

A  one  0    1
          1    5
   two  0    2
          1    6
B  one  0    3
          1    7
   two  0    4
          1    8
dtype: int64

D MultiIndex([(0, 'A', 'one'), (0, 'A', 'two'), (0, 'B', 'one'), (0, 'B', 'two'), (1, 'A', 'one'), (1, 'A', 'two'), (1, 'B', 'one'), (1, 'B', 'two')], names=[None, None, None]) with values [1, 2, 3, 4, 5, 6, 7, 8]

💡 Hint

Stack moves the innermost column level to the row index, creating a Series with a MultiIndex.

Stack converts the innermost column level into the innermost row index level. Here, the second level of columns ('one', 'two') becomes part of the row index, resulting in a Series indexed by the original row index and the first column level.

❓ data_output

intermediate

Unstacking a Series with MultiIndex

What is the output of unstacking the following Series s.unstack(level=1)?

Data Analysis Python

import pandas as pd

index = pd.MultiIndex.from_tuples([(0, 'A'), (0, 'B'), (1, 'A'), (1, 'B')])
data = [10, 20, 30, 40]
s = pd.Series(data, index=index)
result = s.unstack(level=1)

A

   0   1
A  10  30
B  20  40

B

0  A    10
   B    20
1  A    30
   B    40
dtype: int64

C

A  0    10
   1    30
B  0    20
   1    40
dtype: int64

D

   A   B
0  10  20
1  30  40

💡 Hint

Unstack moves the specified level of the row index to columns.

Unstacking level 1 moves the second level of the MultiIndex ('A', 'B') to columns, resulting in a DataFrame with index 0 and 1 and columns 'A' and 'B'.

🧠 Conceptual

advanced

Understanding the effect of stack on DataFrame shape

If a DataFrame has shape (4, 3) and you apply df.stack(), what will be the shape of the resulting object?

A A Series with length 12

B A DataFrame with shape (3, 4)

C A DataFrame with shape (4, 3)

D A Series with length 7

💡 Hint

Stack moves columns into the row index, flattening the DataFrame into a Series.

Stack converts columns into a new inner row index level, so the total number of rows in the resulting Series is rows * columns = 4 * 3 = 12.

🔧 Debug

advanced

Identify the error in unstacking a Series

What error will this code raise and why?

import pandas as pd
s = pd.Series([1, 2, 3], index=[0, 1, 2])
s.unstack()

A ValueError: Index contains duplicate entries, cannot unstack

B IndexError: Too many levels: Index has only 1 level, not 2

C AttributeError: 'Series' object has no attribute 'unstack'

D TypeError: unstack() missing required argument 'level'

💡 Hint

Unstack requires a MultiIndex to move one level to columns.

Unstack tries to pivot one level of a MultiIndex to columns. Here, the Series has a simple Index with only one level, so unstacking fails with IndexError.

🚀 Application

expert

Transforming a DataFrame with stack and unstack to swap index and columns

You have a DataFrame df with shape (3, 4). You want to swap its rows and columns using stack and unstack only (no transpose). Which sequence of operations will achieve this?

A <code>df.stack(level=1).unstack(level=0)</code>

B <code>df.unstack().stack(level=1)</code>

C <code>df.stack().unstack(level=0)</code>

D <code>df.unstack(level=0).stack(level=1)</code>

💡 Hint

Stack moves columns to rows, unstack moves rows to columns. Use level parameters carefully.

Stack moves the innermost column level to the row index, then unstack with level=0 moves the original row index to columns, effectively swapping rows and columns.