OceanBase Database V4.6.0 is a major upgrade for modern data processing. This version significantly enhances OceanBase Database in areas such as AI search, HTAP analysis performance, compatibility and ecosystem integration, security and reliability, and observability and operations.
Key features
Enhanced AI search and vector capabilities
Native SQL hybrid search
The
HYBRID_SEARCHSQL path is now available, supporting hybrid search with vector, full-text, and scalar conditions, and including fusion strategies such as RRF/WRRF. This allows for a more unified SQL approach to hybrid search scenarios, lowering the entry barrier.Enhanced full-text search and hybrid search execution
New Match Phrase (with
slop) and extended full-text index types, with optimized query and build performance. The hybrid search execution framework has been restructured to improve efficiency in full-text paths within hybrid search scenarios. This results in better performance for phrase matching and search in knowledge base and Q&A retrieval scenarios.Search Index
A new Search Index feature is introduced to meet the efficient retrieval needs of complex types such as JSON in hybrid search scenarios. It supports path and type-based index range configuration, enhancing query efficiency.
Enhanced vector database capabilities
Optimized HNSW incremental scenarios, support for IVF on tables without primary keys, and new recall rate assessment capabilities. Sparse vector optimizations further enhance vector search capabilities, making them more observable, optimizable, and operable.
Document AI capabilities
Added document reading, parsing, and splitting capabilities, which can be combined with vector search and hybrid search. This completes the data parsing to retrieval chain for document data, facilitating the development of RAG applications.
Optimized HTAP and analysis performance
Enhanced columnar replica auto-routing and unified access
Supports unified access to columnar replicas and F replicas, with automatic query routing based on policies. This reduces access transformation costs and improves traffic governance flexibility in TP/AP mixed workloads.
Strong-consistency reads for columnar replicas
Added strong-consistency read capabilities to the columnar replica access path. This provides clearer options for analysis queries with higher consistency requirements.
Enhanced parallel execution and incremental query capabilities
Improved PX queuing and task rebalancing mechanisms, and support for incremental Encoding/Skip Index optimization. This ensures more stable and predictable performance for large queries in high-load and skewed data scenarios.
Enhanced SQL/PL execution efficiency and development experience
Continued optimization of query optimizer and executor
Multiple enhancements in hard parsing, predicate derivation, pushdown, and window calculation. This further improves plan generation and execution efficiency for complex SQL, reducing resource consumption.
Enhanced PL engine capabilities and performance
Includes dynamic SQL parameterization, FORALL extension, asynchronous transaction commit, and streaming cursor optimization. Stored procedures and batch processing applications perform better in high-concurrency scenarios.
Enhanced compatibility and ecosystem integration
Enhanced MySQL/Oracle compatibility
Supports MySQL session temporary tables and optimizes Oracle mode global temporary table performance. This reduces migration and dual-mode usage adaptation costs.
Enhanced materialized view capabilities
Expanded incremental refresh scenarios, cascading refresh for nested materialized views, and enhanced MLOG automation management and rewriting capabilities. This makes precomputation maintenance more flexible in data warehouse and reporting scenarios, reducing operational complexity.
Enhanced lakehouse ecosystem integration
Improved Iceberg (including V3 and Time Travel) and REST Catalog capabilities, and enhanced support for Parquet/ORC complex types. This ensures smoother integration with mainstream data lake ecosystems, completing the data analysis chain.
Enhanced security, reliability, and observability
Enhanced data security and authentication
Added column-level sensitive data protection capabilities and support for
caching_sha2_passwordauthentication. This provides finer-grained data protection without changing the main business access method.Enhanced backup verification capabilities
Added backup verification capabilities, allowing pre-recovery checks of backup and archive data availability. This reduces the risk of issues being exposed during the recovery phase, improving disaster recovery process determinism.
Enhanced observability and operational diagnostics
Improved SQLSTAT, ASH, and Plan Cache capabilities, with more detailed wait events and enhanced diagnostic paths. This improves fault localization and performance diagnostics efficiency, making daily operations more controllable.
Window Compaction
Added window compaction capabilities, allowing on-demand scheduling within specified maintenance windows. This enables resource-based control of compaction impact on business operations, facilitating storage and compaction management in large-scale environments.
Enhanced KV and synchronization capabilities
Enhanced OBKV capabilities
Supports weak reads and nearby IDC reads, with improved TTL tasks, HBase monitoring, and hot query optimization. This provides better latency and observability in time series, wide-table, and hot access scenarios.
Enhanced OBCDC capabilities
Supports incremental synchronization after table-level restore and improved compatibility with virtual generated columns. This ensures a more stable data synchronization chain in complex change scenarios, reducing downstream compatibility costs.
References
This article introduces the key new features and capabilities of OceanBase Database V4.6.0. For more details about this version and other versions, please refer to the following documentation: