xushiyan commented on code in PR #14300: URL: https://github.com/apache/hudi/pull/14300#discussion_r2540259382
########## website/releases/release-1.1.0.md: ########## @@ -0,0 +1,366 @@ +--- +title: "Release 1.1.0" +layout: releases +toc: true +--- + +## [Release 1.1.0](https://github.com/apache/hudi/releases/tag/release-1.1.0) + +Apache Hudi 1.1.0 is a major release that brings significant performance improvements, new features, and important changes to the platform. This release focuses on enhanced table format support, improved indexing capabilities, expanded engine support, and modernized record merging APIs. + +## Highlights + +- **Pluggable Table Format Framework** - Native integration of multiple table formats with unified metadata management +- **Spark 4.0 and Flink 2.0 Support** - Full support for latest major compute engine versions +- **Enhanced Indexing** - Non-global Record Index, partition-level bucket index, native HFile writer, and Column Stats V2 +- **Table Services Optimization** - Parquet file stitching and incremental scheduling for compaction/clustering +- **Storage-based Lock Provider** - Multi-writer concurrency control without external dependencies +- **Record Merging Evolution** - Deprecation of payload classes in favor of merge modes and merger APIs + +--- + +## New Features + +### Table Format + +#### Pluggable Table Format Support + +Hudi 1.1.0 introduces a new Pluggable Table Format framework that enables native integration of multiple table formats within the system. This foundation includes a base interface for pluggable table formats, designed to simplify extension and allow seamless interoperability across different storage backends. The Metadata Table (MDT) integration has been enhanced to support pluggability, ensuring modularity and unified metadata management across all supported table formats. + +This release brings native Apache Hudi integration through the new framework, allowing users to leverage Hudi's advanced capabilities directly while maintaining consistent semantics and performance. The configuration `hoodie.table.format` is set to `native` by default, which works as the Hudi table format. **No configuration changes are required** for existing and new Hudi tables. As additional table formats are supported in future releases, users will be able to set this configuration to work natively with other formats. + +#### Table Version 9 with Index Versioning + +Hudi 1.1 introduces table version 9 with support for index versioning. Indexes in the Metadata Table (column stats, secondary index, expression index, etc) now have version tracking. In 1.1, these indexes use V2 layouts with enhanced capabilities including comprehensive logical data type support. Tables migrated from older versions will retain V1 index layouts, while new tables created with 1.1 use V2. Both versions remain backward compatible, and no action is required when upgrading to 1.1. + +### Indexing + +#### Non-Global Record Index + +In addition to the global Record Index introduced in 0.14.0, Hudi 1.1 adds a non-global variant that guarantees uniqueness for partition path and record key pairs. This index speeds up lookups in very large partitioned datasets. Review Comment: made some fixes for using 1.1.0. but not strictly needed. we sometimes say 1.1 to mean 1.1.x because patch releases is only meant for minor fixes so behavior and feature-wise, all 1.1.x should be the same. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
