Hi everyone, We are pleased to announce the release of Apache Beam version 2.67.0.
This release brings several improvements, new features, and important fixes. You can download the release here: https://beam.apache.org/get-started/downloads/ Key Highlights: Python: Streaming writes across IOs (IOBase/FileBasedSink; Avro/Parquet/Text/TFRecords). Java: Debezium IO upgraded to 3.1.1 (requires Java 17). Beam SQL: DATABASE support (Iceberg implementation). Python: Milvus enrichment handler (vector/keyword/hybrid search). Spark Classic: Processing-time timers. Java: JdbcIO.WriteWithResults now respects BatchSize/MaxBufferingDuration. Beam YAML: BigTableWrite connector (with mutation). This release includes breaking changes such as Go PubsubIO Read taking value ReadOptions (requires one of Topic or Subscription), and SQL ParquetTable requiring LOCATION directories to end with slash. Please review the full release notes carefully. For detailed information on all changes, improvements, breaking changes, and bug fixes, please see the full release notes: https://github.com/apache/beam/releases/tag/v2.67.0 Thank you to everyone who contributed to this release! Best regards, The Apache Beam Team
