Milestones
- The new SQL parsing module removes the dependence on Druid. We only need to parse the sharding context, and adopt a "semi understanding" concept for SQL to further improve performance and compatibility, and reduce code complexity
- The new SQL rewrite module adds an optimized rewrite module
- The new SQL merge module is rebuilt into 3 merging engines: streaming, memory and decorator
New Features
- Support Oracle, SQLServer and PostgreSQL
- Non functional sub-query support
Enhancements
- ISSUE #256 Configurable display sharding execute to SQL log
- ISSUE #291 Processing SQL use stream mode that contains only GroupBy
Changes
- Simplify the distributed self increasing sequence. Each table is supported by simplifying a multiple self increasing sequence to a single table supporting only a single distributed self increasing sequence, and no longer supporting workerID settings through environment variables
- Remove support for OR
Bug Fixes
- ISSUE #239 LIMIT routes to multiple query result sets. If there is only one result set that is not empty, the result of paging is incorrect
- ISSUE #263 Sharding and logical table configuration can be case insensitive
- ISSUE #292 When the memory mode handles GROUP BY statement, if there is paging information, it needs to be rewritten
- ISSUE #295 LIMIT 0 does not filter the result set according to paging restrictions