Deduplicate DataFrames by Subset Columns in Pandas

Question

Given an event log with `user_id`, `event_name`, and `timestamp`, remove duplicate user actions on the same event, keeping only the most recent occurrence (`keep='last'`).

Accepted Answer

Optimal Solution:
import pandas as pd

# Sort by timestamp ascending first, then keep='last'
deduped_df = df.sort_values('timestamp').drop_duplicates(
    subset=['user_id', 'event_name'],
    keep='last'
).reset_index(drop=True)

Technical Explanation:
`drop_duplicates(subset=[...], keep='last')` inspects the specified columns and retains only the last row among duplicates.

user_id	event_name	timestamp
101	add_to_cart	2026-02-01 10:00:00
101	add_to_cart	2026-02-01 10:05:00
102	add_to_cart	2026-02-01 11:00:00

user_id	event_name	timestamp
101	add_to_cart	2026-02-01 10:05:00
102	add_to_cart	2026-02-01 11:00:00