Iceberg Integration
- Advanced Snapshot Management
- Support for Branches & Tags
- Support for Positional Deletes
- Support for COW & MOR modes
- Vectorised Reads & Writes
- Support for queries with Partition clause
- Support for LOAD DATA queries
- Better Statistics Management
Compaction
- Rebalance compaction (Hive ACID)
- Compaction requests prioritization (compaction pooling)
- Iceberg compaction (Major)
Hive ACID
- Use sequences for TXN_ID generation (performance)
- Read-only transactions optimization
- Zero-wait readers
- Optimistic and Pessimistic concurrency control
- Lockless reads
Hive Metastore
- API optimization (performance)
- Dynamic leader election
- HMS support for Thrift over HTTP
- JWT authentication for Thrift over HTTP
- HMS metadata summary
Hive Replication
- Optimised Bootstrap Solution
- Support for Snapshot Based Replication for External Table
- Better Replication Tracking Metrics
- Support for Checkpointing during Replication
Security
- Authorizations in alter table/view, UDFs, and Views created from Apache Spark.
- Authorizations on tables created based on storage handlers.
- Critical CVE fixes of transitive dependencies.
Compiler Side Improvements
- Materialized view support for Iceberg tables
- Improvements to refresh materialized views
- Date/Timestamp fixes and improvements
- Anti join support
- Split update support
- Branch pruning
- Column histogram statistics support
- Calcite upgrade to 1.25
- HPL/SQL improvements
Miscl.
- Support for ESRI GeopSpatial UDF's
- Added support for Apache Ozone
- Support Hadoop-3.3.6
- Supports Tez 0.10.3
- Works with Aarch64 (ARM)