Top 5 Trending Products in the Last 7 Days

Problem

Given a stream of purchase events (productId, timestamp), at any point return the top 5 most purchased products in the last 7 days (sliding window).

Example

events = [(P1,t1),(P2,t2),(P1,t3),...]
query at t_now → top 5 products by count in [t_now-7days, t_now]

Constraints

• Up to 10^7 events
• Up to 10^6 unique products
• Queries happen in real time

Approaches

1. Offline batch: aggregate per day, maintain 7-day rolling sum
2. Online: bucketed counters (1 bucket per hour) + min-heap for top-k

Myntra follow-up

Extend to return trending products per category and per city simultaneously.