databricks Platform · Updated Jun 2, 2026

Incremental refresh for materialized views in Databricks SQL

Databricks SQL materialized views now support incremental refresh in Public Preview.

Public Preview Open docs ↗release note ↗

What it does

Databricks SQL materialized views now support incremental refresh in Public Preview. Instead of recomputing the entire result set on every refresh, the engine detects changes in base tables and updates only the affected rows, dramatically reducing refresh cost and latency for large aggregations. Highlights: Incremental refresh for a broad set of SQL constructs including joins and aggregations.

Why it matters

Instead of recomputing the entire result set on every refresh, the engine detects changes in base tables and updates only the affected rows, dramatically reducing refresh cost and latency for large aggregations.

Key capabilities

Incremental refresh for a broad

Incremental refresh for a broad set of SQL constructs including joins and aggregations.

Automatic fallback to full refresh

Automatic fallback to full refresh when incremental is not possible.

Refresh scheduling and monitoring in

Refresh scheduling and monitoring in the SQL editor.

Prerequisites

A SQL warehouse on the current channel.

Base tables stored in Delta with change data feed available.

Limitations · remember

!Some non-deterministic functions force a full refresh.

!Incremental refresh statistics are not yet exposed in system tables.

Architecture

Databricks Platform → Platform

In short

Read the full documentation ↗

Key takeaway

Databricks SQL materialized views now support incremental refresh in Public Preview.

2026-06-02-incremental-refresh-for-materialized-views-in-databricks-sql Generated by the Databricks Updates Agent · browse all updates