GroupBy performance considerations in Pandas - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of GroupBy with multiple aggregations

What is the output of this code snippet using pandas GroupBy with multiple aggregations?

Pandas

import pandas as pd

df = pd.DataFrame({
    'Category': ['A', 'A', 'B', 'B', 'C'],
    'Value': [10, 20, 10, 30, 40]
})
result = df.groupby('Category').agg({'Value': ['sum', 'mean']})
print(result)

A

Category  Value
sum       30
mean      15.0

B

          Value     
           sum  mean
Category           
A           30  15.0
B           40  20.0
C           40  40.0

C

Value
sum    100
mean    20.0

D

Category  Value
A        30
B        40
C        40

💡 Hint

Remember that multiple aggregations create a MultiIndex column in the result.

When using multiple aggregation functions on a column, pandas returns a DataFrame with MultiIndex columns showing each aggregation result per group.

❓ data_output

intermediate

Number of groups created by GroupBy

Given this DataFrame, how many groups will pandas GroupBy create when grouping by 'Type'?

Pandas

import pandas as pd

df = pd.DataFrame({
    'Type': ['X', 'Y', 'X', 'Z', 'Y', 'X'],
    'Score': [5, 10, 15, 20, 25, 30]
})
groups = df.groupby('Type')
print(len(groups))

A 3

B 6

C 4

D 1

💡 Hint

Count unique values in the 'Type' column.

GroupBy creates one group per unique value in the grouping column. Here, 'X', 'Y', and 'Z' are unique, so 3 groups.

🔧 Debug

advanced

Identify the cause of slow GroupBy operation

This code runs very slowly on a large DataFrame. What is the main reason for the slow performance?

Pandas

import pandas as pd
import numpy as np

df = pd.DataFrame({
    'Category': np.random.choice(['A', 'B', 'C', 'D'], size=10_000_000),
    'Value': np.random.rand(10_000_000)
})

result = df.groupby('Category').apply(lambda x: x['Value'].sum())

A Using apply with a lambda function instead of built-in aggregation slows down performance.

B Grouping by a categorical column always causes slowdowns.

C The DataFrame is too small to benefit from GroupBy optimizations.

D The random values in 'Value' column cause the slowdown.

💡 Hint

Built-in aggregation functions are faster than apply with custom functions.

Using apply with a lambda function prevents pandas from using optimized C code for aggregation, causing slow performance on large data.

🧠 Conceptual

advanced

Effect of sorting on GroupBy performance

How does setting the 'sort' parameter to False in pandas GroupBy affect performance and output?

A Setting sort=False removes duplicate groups, improving performance.

B Setting sort=False causes an error because sorting is mandatory.

C Setting sort=False sorts groups alphabetically, slowing down performance.

D Setting sort=False improves performance by skipping sorting but keeps groups in original order.

💡 Hint

Sorting groups is optional and can be skipped for speed.

By default, GroupBy sorts group keys which adds overhead. Setting sort=False skips sorting, speeding up grouping and preserving original order.

🚀 Application

expert

Optimizing memory usage in GroupBy with large categorical data

You have a DataFrame with 50 million rows and a 'Category' column with 100 unique values. Which approach best optimizes memory and performance for grouping?

A Convert 'Category' to string dtype before grouping.

B Keep 'Category' as object dtype and group directly.

C Convert 'Category' to pandas Categorical dtype before grouping.

D Drop the 'Category' column before grouping to save memory.

💡 Hint

Categorical dtype uses less memory and speeds up grouping on repeated values.

Converting to Categorical reduces memory by storing codes instead of strings and speeds up groupby operations on large data with repeated categories.