Python data engineering interview problem. Difficulty: intermediate. Pattern: Dictionaries. About 16 minutes. Part of the Pro drill bank.
Keep each driver's recent trail: drop repeated pings and pings more than 30 seconds older than that driver's newest ping. Treat this as a production helper: match the contracted return shape, including empty and duplicate inputs.
Implement latest_locations(pings: list[dict]) -> dict. A ping is {"driver", "ts", "lat", "lon"} with ts in whole seconds. Pings arrive in any order. For every driver, find the newest ts seen for that driver. Keep a ping only if it is not a repeat of an identical ping (same driver, ts, lat and lon) and it is at most 30 seconds older than that driver's newest ping (a ping exactly 30 seconds older is kept). Return {driver: [(ts, lat, lon), ...]} with each driver's pings sorted by ts and drivers sorted by name. The last tuple of a driver is their latest location.
Input: latest_locations([{"driver": "d1", "ts": 150, "lat": 1.2, "lon": 3.4}, {"driver": "d1", "ts": 100, "lat": 1.0, "lon": 3.0}, {"driver": "d1", "ts": 155, "lat": 1.3, "lon": 3.5}, {"driver": "d1", "ts": 150, "lat": 1.2, "lon": 3.4}, {"driver": "d2", "ts": 40, "lat": 9.0, "lon": 9.5}]) Output: {'d1': [(150, 1.2, 3.4), (155, 1.3, 3.5)], 'd2': [(40, 9.0, 9.5)]} d1's newest ping is 155, so its ping at 100 is too old and the repeated ping at 150 appears once. d2 has its own cutoff, so its only ping stays.
Topics: lakebench, python, streaming, dedupe, late data.
More Python interview questions · All interview problems · Learn data engineering
Interview-style drill: Keep each driver's recent trail: drop repeated pings and pings more than 30 seconds older than that driver's newest ping.
Implement `latest_locations(pings: list[dict]) -> dict`. A ping is `{"driver", "ts", "lat", "lon"}` with `ts` in whole seconds. Pings arrive in any order. For every driver, find the newest `ts` seen for that driver. Keep a ping only if it is not a repeat of an identical ping (same driver, ts, lat and lon) and it is at most 30 seconds older than that driver's newest ping (a ping exactly 30 seconds older is kept). Return `{driver: [(ts, lat, lon), ...]}` with each driver's pings sorted by `ts` and drivers sorted by name. The last tuple of a driver is their latest location.