Select Top N Records Per Group with nlargest() in Pandas

Question

Given an e-commerce catalog, select the top 2 highest-priced products for each `department`.

Accepted Answer

Optimal Solution:
import pandas as pd

# Method 1: Using rank (Vectorized & Fast)
df['rank'] = df.groupby('department')['price'].rank(method='first', ascending=False)
top_2_df = df[df['rank'] <= 2].drop(columns='rank')

# Method 2: Using nlargest
# top_2_df = df.groupby('department', group_keys=False).apply(lambda g: g.nlargest(2, 'price'))

Technical Explanation:
Ranking with `groupby('department')['price'].rank(method='first', ascending=False)` allows direct filtering `<= 2` in fast vectorized C-speed.

product_id	department	price
1	Tech	1200
2	Tech	800
3	Tech	400
4	Books	35

department	product_id	price
Tech	1	1200
Tech	2	800
Books	4	35