Study Databricks DA-ASSOC Views and Time Travel: key concepts, common traps, and exam decision cues.
Databricks often tests object selection through scenarios that sound similar on the surface. The right answer depends on whether the need is repeated read performance, continuously arriving data, governed storage control, security filtering, or historical access to older table versions.
| Need | Better choice |
|---|---|
| repeated analytics on a precomputed result | materialized view |
| continuously arriving data stream | streaming table |
| Databricks-managed storage lifecycle | managed table |
| externally managed storage path | external table |
| row or column restriction for different users through SQL logic | dynamic view |
| query an older Delta table version | time travel |
| Pair | Keep this distinction clear |
|---|---|
| materialized view vs streaming table | repeated query acceleration versus continuous data pipeline pattern |
| managed table vs external table | Databricks-managed lifecycle versus externally managed storage |
| dynamic view vs materialized view | security or conditional exposure logic versus stored precomputed result |
| Trap | Better rule |
|---|---|
| picking materialized view for real-time arriving sensor data | streaming table usually fits ongoing arrival better |
| assuming time travel always works forever | retention and cleanup behavior still matter |
| treating managed versus external as a visualization decision | it is a storage and lifecycle boundary decision |
This objective is mainly about choosing the right object for the data-change pattern. If data arrives continuously and should stay current, think streaming table. If the requirement is historical inspection of a Delta table, think time travel. The exam usually rewards matching refresh semantics and historical access needs to the right feature.