A data analyst is building a reporting layer in Databricks SQL. The business requires a table that always reflects the latest data from multiple upstream tables but should not re-execute expensive join logic on every query. Which Databricks SQL object best meets this requirement?

Question

Accepted Answer

B. Materialized view. A materialized view in Databricks SQL pre-computes and stores the results of the defining query (including complex joins) so that queries against it read from stored results rather than re-executing the join logic. Databricks refreshes the materialized view incrementally to keep it current with upstream changes. Standard views re-execute on every query (expensive for complex joins). External Parquet tables are static and require manual refresh. Delta Live Tables streaming tables are for streaming ingestion, not pre-computed join results.

A data analyst is building a reporting layer in Databricks SQL. The business requires a table that always reflects the latest data from multiple upstream tables but should not re-execute expensive join logic on every query. Which Databricks SQL object best meets this requirement?

Related Questions

Discussion