DataDriven
LearnPracticeInterviewDiscussDailyJobs

The current daily_revenue pipeline computes today only, runs once a night, and freezes each day's nu

A medium Pipeline Design interview practice problem on DataDriven. Write and execute real pipeline design code with instant grading.

Domain
Pipeline Design
Difficulty
medium

Problem

The current daily_revenue pipeline computes today only, runs once a night, and freezes each day's number forever. A late mobile event from five days ago arrives today and lands nowhere because the pipeline never looks back. The section's simplest workable fix is the seven-day rerun window: every morning, recompute the last seven days, not just today. The pattern requires three things: an idempotent write (re-running produces the same answer, not duplicates), partition-level overwrite by event_date (yesterday's row gets replaced, not added to), and source data retained for at least the window length. Set the rerun window: add an orchestrator that schedules a daily run with a seven-day rerun window in its name, and replace the destination with a lakehouse format whose partition overwrite makes the rerun idempotent.

Practice This Problem

Solve this Pipeline Design problem with real code execution. DataDriven runs your solution and grades it automatically.

Related

  • All Practice Problems
  • Mock Interview Mode
  • System Design Interview Questions
  • Data Engineering Interview Prep Guide
  • Daily Challenge
  • Data Engineering Lessons