databricks-table-valued-functions

A Databricks table-valued function (TVF) is used to build reusable, parameterized SQL queries for the semantic layer, enabling natural language data access through Genie Spaces. It encapsulates business logic so users can query data consistently without writing raw SQL.

You create parameterized queries for Genie Spaces by building table-valued functions with proper parameter typing, such as using STRING for dates. This Skill guides the end-to-end process including schema validation, null safety, and Asset Bundle deployment patterns.

Common SQL errors when building Databricks TVFs include cartesian products, incorrect parameter ordering, and LIMIT clause issues. Proper parameter typing and following SQL best practices for null safety prevent these Genie Space compatibility problems.

Yes, you can deploy Databricks table-valued functions using Asset Bundles. This Skill provides deployment patterns and validation steps to ensure TVFs are correctly packaged and optimized for natural language querying within Genie Spaces.

You should use a TVF instead of a standard SQL view when you need parameterized queries for dynamic filtering in Genie Spaces. TVFs support parameter typing and null safety, which are required for natural language data access and reusable semantic layer logic.