Metric view feature availability

Metric views were introduced in Databricks Runtime 16.4. The following table lists minimum Databricks Runtime version requirements for specific features added in later versions.

Databricks Runtime version	Features
18.1	Window measure offset modifier (YAML spec 1.1) inclusive and exclusive modifiers on trailing and leading window ranges (YAML spec 1.1)
18.0	BI compatibility mode REFRESH MATERIALIZED VIEW
17.3	Snowflake schema Agent metadata — comments, synonyms, formatting, display names (YAML spec 1.1) TEMPORARY metric views Materialization (Experimental) JDBC/ODBC support through Thrift server Join optimization with rely.at_most_one_match

YAML specification versions

The metric view YAML specification is versioned independently of Databricks Runtime. The following sections describe what each version adds.

Version 1.1

Adds support for agent metadata (display names, formatting, synonyms) and the optional comment field for metric views, fields, and measures. See Agent metadata in metric views.

Also adds the following window-measure fields and modifiers (require Databricks Runtime 18.1 or newer):

• The offset field on window specifications. See How offset shifts the window frame.
• The inclusive and exclusive modifiers on trailing and leading range values. See Include or exclude the anchor row.

Version 0.1

Initial release of the metric view YAML specification.