← Back
Databricks
Databricks Serverless Environment Version 5 now available; adds SQL scripting and streaming improvements
· releasefeatureplatformperformance · docs.databricks.com ↗

Overview

Databricks has released Serverless environment version 5, corresponding to Serverless version 18.0 (February 27, 2026). This release brings significant enhancements to SQL capabilities, streaming queries, and data integration features for serverless compute environments.

New Features

SQL and Query Enhancements

  • SQL scripting is now generally available: SQL scripting moves from preview to general availability, enabling complex procedural SQL workflows.
  • Parameter markers everywhere: Named (:param) and unnamed (?) parameter markers are now supported virtually anywhere a literal value can be used, including DDL statements and column type definitions.
  • IDENTIFIER clause expansion: The IDENTIFIER clause now works in nearly all places an identifier is permitted, including column aliases and column definitions.
  • Literal string coalescing: Sequential string literals (e.g., 'Hello' ' World') can now be coalesced into single values in any context where string literals are allowed.
  • SQL window functions in metric views: Window functions for running totals, rankings, and other calculations are now supported in metric views.

Streaming and Performance

  • Dynamic shuffle partition adjustment: Stateless streaming queries can now adjust the number of shuffle partitions without requiring a restart.
  • Adaptive Query Execution (AQE) and auto-optimized shuffle (AOS): Both features are now supported in stateless streaming queries for improved performance.

Python UDFs and Integrations

  • Shared isolation execution environment for Unity Catalog Python UDFs: Python UDFs with the same owner can now share an isolation environment by default, improving performance and reducing memory usage. Use the STRICT ISOLATION clause to force complete isolation when needed.
  • Redshift JDBC driver upgraded to 2.1.0.28: Better compatibility and support for Amazon Redshift integrations.

New Functions

  • BITMAP_AND_AGG function: New aggregate function for bitmap operations.
  • Theta sketch functions: New library for approximate distinct count and set operations, including theta_sketch_agg, theta_union_agg, and theta_intersection_agg.

What Developers Need to Know

All serverless users receive automatic updates to the latest version. Environment versions provide a stable client API based on Spark Connect, ensuring application compatibility while allowing independent server upgrades. Existing workloads do not require code changes to benefit from performance improvements, security enhancements, and bug fixes in this release.