LLearningPath
Executing queries using Databricks SQL and SQL Warehouses

20% domain · module 3

Views, materialized views, streaming tables, time travel

Starts at 1:22:18 in this video — SQL Warehouse full course — streaming tables (1:22:18)

Pick the object by freshness vs reuse

  • View (dynamic) — does not store data; reruns the query; light logic, always current
  • Materialized view — stores a refreshed result; heavy BI queries on relatively stable/gold data
  • Streaming table — incremental storage; continuous arrivals (sensors, files via Auto Loader)
  • Delta table — source of truth with time travel

Exam pairing: streaming table for continuously arriving sensor data; materialized view for frequent complex queries on (mostly) static BI data. Do not invert them.

Time travel

sqlSELECT * FROM catalog.gold.orders VERSION AS OF 12;
SELECT * FROM catalog.gold.orders TIMESTAMP AS OF '2026-08-01';

If a 30-day-old version errors, the usual reason is VACUUM (or retention) deleted the data/log files for that version — not a rename and not a permission that “disables time travel.”

Views · Materialized views · Time travel

Next module →