Event files like the one below arrive continuously in a UC volume, and the team starts Auto Loader with schema inference (no explicit schema, cloudFiles.schemaLocation set): ```json {"user_id": 1042, "event_ts": "2026-05-01T10:22:31Z", "amount": 49.9, "profile": {"tier": "gold", "region": "emea"}} ``` What schema does Auto Loader infer by default, and how can the team get typed columns such as numbers and timestamps?

1. A. All fields — including nested ones — are inferred as string; set cloudFiles.inferColumnTypes to true to infer types like long, double, and timestamp ✓ A
2. B. user_id becomes long, amount becomes double, event_ts becomes timestamp, and profile becomes a struct, because Auto Loader always samples rows and picks the narrowest matching type 💬 B
3. C. The profile object is rejected, because Auto Loader cannot infer nested structures from JSON files 💬 C
4. D. Top-level fields are strings while nested fields keep their JSON types, so only profile.tier and profile.region are correctly typed 💬 D

Reveal answer

Correct answer: All fields — including nested ones — are inferred as string; set cloudFiles.inferColumnTypes to true to infer types like long, double, and timestamp

Reference: https://docs.databricks.com/aws/en/ingestion/cloud-object-storage/auto-loader/