filter() for group-level filtering in Pandas - Time & Space Complexity

Time Complexity: filter() for group-level filtering

O(n)

Understanding Time Complexity

We want to understand how the time needed changes when we use filter() on groups in pandas.

Specifically, how does the work grow as the data gets bigger?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import pandas as pd

df = pd.DataFrame({
    'Category': ['A', 'A', 'B', 'B', 'C', 'C'],
    'Value': [10, 15, 10, 5, 20, 25]
})

grouped = df.groupby('Category')
filtered = grouped.filter(lambda x: x['Value'].mean() > 12)

This code groups data by 'Category' and keeps only groups where the average 'Value' is more than 12.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: The filter() applies a function to each group, calculating the mean of 'Value' in that group.
• How many times: Once for each group in the data.

How Execution Grows With Input

As the number of rows grows, the number of groups and their sizes affect the work done.

Input Size (n)	Approx. Operations
10	About 10 operations to check values in groups
100	About 100 operations, since each row is checked once in its group
1000	About 1000 operations, scaling with total rows

Pattern observation: The work grows roughly in direct proportion to the total number of rows.

Final Time Complexity

Time Complexity: O(n)

This means the time needed grows linearly with the number of rows in the data.

Common Mistake

[X] Wrong: "Filtering groups with filter() is constant time regardless of data size."

[OK] Correct: The function runs on each group and each row inside, so more data means more work.

Interview Connect

Knowing how group filtering scales helps you explain your choices clearly and shows you understand data size impact.

Self-Check

"What if the filtering function was more complex, like calculating median instead of mean? How would the time complexity change?"

Practice

1. What does the filter() function do when applied to a grouped pandas DataFrame?

easy

A. Sorts groups by their size

B. Filters rows inside each group individually

C. Keeps or removes entire groups based on a condition

D. Aggregates data by calculating group means

Solution

1. Step 1: Understand the purpose of filter() on groups

   The filter() function is used on grouped data to keep or remove whole groups based on a condition applied to the group as a whole.

2. Step 2: Differentiate from row-level filtering

   Unlike filtering rows inside groups, filter() returns entire groups only if the condition is True for that group.

3. Final Answer:

   Keeps or removes entire groups based on a condition -> Option C

4. Quick Check:

   Group-level filtering = Keeps/removes groups [OK]

Hint: Remember: filter() keeps or drops whole groups [OK]

Common Mistakes:

• Thinking filter() filters rows inside groups
• Confusing filter() with aggregation functions
• Assuming filter() sorts groups

2. Which of the following is the correct syntax to filter groups in a pandas DataFrame df grouped by column 'Category' to keep groups with more than 3 rows?

easy

A. df.groupby('Category').filter(lambda x: len(x) > 3)

B. df.filter(groupby='Category', condition=lambda x: len(x) > 3)

C. df.groupby('Category').apply(filter(len(x) > 3))

D. df.filter(lambda x: x['Category'] > 3)

Solution

1. Step 1: Recall correct filter() syntax on grouped data

   The correct syntax is to group by a column, then call filter() with a function that takes each group and returns True or False.

2. Step 2: Check each option

   df.groupby('Category').filter(lambda x: len(x) > 3) correctly groups by 'Category' and filters groups with length > 3. Options A, B, and D use incorrect syntax or wrong method calls.

3. Final Answer:

   df.groupby('Category').filter(lambda x: len(x) > 3) -> Option A

4. Quick Check:

   Correct syntax = groupby + filter(lambda x: condition) [OK]

Hint: Use groupby().filter(lambda x: condition) for group filtering [OK]

Common Mistakes:

• Using filter() without groupby()
• Passing wrong arguments to filter()
• Confusing apply() with filter()

3. Given the DataFrame df below:

import pandas as pd
data = {'Team': ['A', 'A', 'B', 'B', 'B', 'C'], 'Points': [10, 15, 10, 5, 7, 20]}
df = pd.DataFrame(data)
filtered = df.groupby('Team').filter(lambda x: x['Points'].sum() > 20)
print(filtered)

What will be the output?

medium

A. Rows where Team is 'A' and 'C'

B. Rows where Team is 'A' and 'B'

C. Rows where Team is 'B' only

D. All rows

Solution

1. Step 1: Calculate sum of Points per Team

   Team A: 10 + 15 = 25, Team B: 10 + 5 + 7 = 22, Team C: 20

2. Step 2: Apply filter condition sum > 20

   Teams A and B have sums > 20, Team C has 20 (not > 20), so only A and B groups pass.

3. Step 3: Check printed rows

   Filtered DataFrame includes rows from Teams A and B only.

4. Final Answer:

   Rows where Team is 'A' and 'B' -> Option B

5. Quick Check:

   Sum > 20 keeps Teams A and B [OK]

Hint: Sum group values, keep groups where sum > 20 [OK]

Common Mistakes:

• Forgetting to include Team A rows
• Including Team C because sum equals 20
• Confusing sum with mean

4. Consider this code snippet:

df.groupby('Category').filter(lambda x: x['Value'] > 10)

Why will this code raise an error?

medium

A. Because the lambda returns a Series, not a single True/False value

B. Because filter() cannot be used after groupby()

C. Because 'Value' column does not exist

D. Because filter() requires a list, not a lambda

Solution

1. Step 1: Understand filter() expects a boolean scalar per group

   The function inside filter() must return a single True or False for the whole group.

2. Step 2: Analyze the lambda function

   The lambda returns a Series of booleans (x['Value'] > 10) for each row, not a single boolean.

3. Final Answer:

   Because the lambda returns a Series, not a single True/False value -> Option A

4. Quick Check:

   filter() needs single True/False per group [OK]

Hint: Lambda must return one True/False per group, not Series [OK]

Common Mistakes:

• Returning a Series instead of a scalar boolean
• Thinking filter() works like row-level filter
• Misunderstanding filter() usage after groupby

5. You have a sales DataFrame df with columns 'Store', 'Sales', and 'Month'. You want to keep only stores where the average monthly sales are above 5000. Which code correctly filters the DataFrame using filter()?

hard

A. df.filter(lambda x: x['Sales'].mean() > 5000, groupby='Store')

B. df.groupby('Store').filter(lambda x: x['Sales'] > 5000)

C. df.groupby('Store').apply(lambda x: x['Sales'].mean() > 5000)

D. df.groupby('Store').filter(lambda x: x['Sales'].mean() > 5000)

Solution

1. Step 1: Identify correct use of filter() with groupby

   We must group by 'Store' and use filter() with a lambda returning True if the group's average sales > 5000.

2. Step 2: Check each option

   df.groupby('Store').filter(lambda x: x['Sales'].mean() > 5000) correctly uses filter() with lambda x: x['Sales'].mean() > 5000. df.filter(lambda x: x['Sales'].mean() > 5000, groupby='Store') has wrong syntax, C uses apply which returns a Series, and D compares sales values directly, not the mean.

3. Final Answer:

   df.groupby('Store').filter(lambda x: x['Sales'].mean() > 5000) -> Option D

4. Quick Check:

   Use groupby + filter with mean condition [OK]

Hint: Filter groups by mean with groupby().filter(lambda x: x['col'].mean() > val) [OK]

Common Mistakes:

• Using filter() without groupby()
• Comparing individual sales instead of group mean
• Using apply() instead of filter() for filtering groups