ReleaseBytes
databricks Databricks Release Notes ·

Databricks Unity Catalog Python UDFs (Scalar & Batch) are Now Generally Available

datadatabricksgaengineerdatabricks-unity-catalog
feature announcement

Databricks has announced the general availability of Scalar and Batch Python UDFs within Unity Catalog across all supported compute types. This release enhances UDF functionality by introducing named handlers for scalar UDFs, removing previous call limits, and aligning TIMESTAMP input behavior with PySpark. Users can now also install custom dependencies and securely access Unity Catalog credentials and secrets directly from their Python functions. Certain advanced capabilities, including scalar UDFs on serverless compute, require explicitly setting `environment_version` to 6 or higher.

  • →General Availability of Scalar and Batch Python UDFs in Unity Catalog
  • →Enhanced Capabilities for Unity Catalog Python UDFs
Features (1) ›
  • General Availability of Scalar and Batch Python UDFs in Unity Catalog

    Databricks has made Scalar and Batch Python UDFs generally available across all supported compute types within Unity Catalog. This release includes named handlers for row-at-a-time scalar UDFs and removes the five-UDF-call-per-query limit, improving flexibility and performance. It also aligns scalar TIMESTAMP inputs with PySpark behavior for consistency.

Enhancements (1) ›
  • Enhanced Capabilities for Unity Catalog Python UDFs

    Python UDFs now support installing custom dependencies directly and provide access to Unity Catalog service credentials and secrets, enabling more robust and integrated functions. Some of these advanced capabilities, including newly available scalar features on serverless compute, require setting the UDF definition's `environment_version` to 6 or above.

Read the original announcement →

https://docs.databricks.com/aws/en/release-notes/product/2026/september#scalar-and-batch-unity-catalog-python-udfs-are-now-generally-available

Related releases