What is a materialized view, and how can it speed up Fabric queries?

• A. A materialized view stores the results of a query logically without materializing data.
• B. A materialized view stores the results of a query physically; periodic refresh keeps it up to date, speeding up repeated analytics workloads.
• C. A materialized view is a live, on-the-fly computation of a query every time.
• D. A materialized view duplicates data to create backups.

Answer: (B)

A materialized view is a result set that is stored physically on disk. Instead of running the underlying query every time, Fabric can read the precomputed, stored results, which makes repeated analytics much faster. The stored results are kept up to date by a periodic refresh, so the data reflects changes in the source tables while balancing freshness and performance. This approach is especially helpful for expensive operations like large joins or aggregations that are frequently queried.

If the data changes between refreshes, the results can become slightly stale, and you pay with storage and a refresh overhead. The other descriptions don’t match how a materialized view works: it’s not just a logical, on-the-fly result, it’s a stored, refreshable snapshot used to speed queries, and it isn’t primarily for backups.