Unity Catalog metric views

Metric views are the core implementation of Unity Catalog business semantics in Unity Catalog. They provide a centralized way to define and manage business metrics by separating measure definitions from dimension groupings, so you can define metrics once and query them across any dimension at runtime.

Standard views lock in aggregations and dimensions at creation time. With a metric view, you define the metric once — for example, sum of revenue divided by distinct customer count — and users can group by any available dimension. The query engine generates the correct computation.

Get started

Use the following pages to create, query, and manage metric views.

• • Create and edit metric views
  • Define metric views using SQL DDL or the Catalog Explorer UI, with built-in YAML validation.
• • Query metric views
  • Query metric views from SQL editors, notebooks, dashboards, Genie spaces, and alerts.
• • Manage metric views
  • Control access, enable collaborative editing, and manage the metric view lifecycle.
• • Tutorial: Build a complete metric view with joins
  • Walk through a complete sales analytics metric view with joins, measures, and agent metadata.

Core features

Use the following pages to learn about modeling, materialization, and agent metadata.

• • Model metric views
  • Define sources, dimensions, measures, and filters. Model relationships using star and snowflake schemas with multi-level joins.
• • Advanced techniques for metric views
  • Build complex metrics using composability and window measures for trailing averages, period-over-period changes, and cumulative totals.
• • Materialization for metric views
  • Pre-compute and incrementally refresh aggregations. The query engine automatically rewrites queries to use materialized views when appropriate. Experimental.
• • Agent metadata in metric views
  • Add synonyms, display names, and formatting rules to improve AI agent accuracy and ensure consistent metric display across tools.
• • Metric view YAML syntax reference
  • Complete reference for the metric view YAML specification, including all fields, types, and formatting examples.