Introducing Apache Spark 4.2
Databricks, Thursday, July 16th, 2026
Apache Spark 4.2 adds metric views, expanded Spark Connect, and SQL primitives for vector and geospatial analytics.
Apache Spark 4.2 introduces metric views that provide governed business definitions, enabling consistent analytics across SQL, BI tools, and AI applications.
The release expands Spark Connect and PySpark capabilities, making it simpler to invoke Spark as a service from a wide range of applications and environments.
New SQL primitives for vector similarity search, geospatial analytics, and time-series analysis push what Databricks describes as AI-native analytics directly into the engine.
Rounding out the release, Auto CDC, Data Source V2 enhancements, and Real-Time Mode improvements simplify processing of continuously changing data with better reliability and lower latency.