Materialized View

A materialized view is an entity that contains both computation logic and data. You can use it for query acceleration, lightweight ETL modeling, and lakehouse federated query acceleration. Start by understanding the concepts, then choose between synchronous and asynchronous materialized views based on your freshness requirements.

Concept Introduction

Materialized View Overview

Understand the applicable scenarios of materialized views and make selection decisions based on freshness, SQL pattern, and refresh method.

Synchronous Materialized View

Synchronous Materialized View

Use synchronous materialized views to keep data strongly consistent with the base table, accelerating real-time aggregation and sorting scenarios.

Transparent Rewriting with Synchronous Materialized View

Learn how the optimizer transparently rewrites queries to synchronous materialized views, including supported patterns and limitations.

Asynchronous Materialized View

Asynchronous Materialized View Overview

Master the refresh modes, multi-table support, partition incremental refresh, and typical use cases of asynchronous materialized views.

Creation and Maintenance

Operations manual for creating asynchronous materialized views, querying them directly, managing refreshes, and maintaining partitions.

Usage Guide

Best practices for scenario judgment, usage principles, refresh strategy selection, and build implementation.

Transparent Rewriting with Asynchronous Materialized View

How the SPJG-pattern-based rewriting algorithm transparently rewrites existing queries to asynchronous materialized views without modifying SQL.

FAQ

Answers to frequently asked questions about partition refresh, base table versions, and operational troubleshooting.