Features
- Enable system table all_manifests for Iceberg tables (#60279)
- Support bucket-domain-name for OSS (#59755)
- Support flexible partial update for routine load (#60128)
- Implement expire_snapshots procedure for Iceberg tables (#59979)
- Support PREVIOUS_DAY function (#60680)
- Support 3 spatial functions: ST_Distance, ST_GeometryType, ST_Length (#60170)
- Support more INTERVAL time-unit (#60347)
- Add OzoneProperties to support Apache Ozone (#60809)
- Support MaxCompute ram_role_arn and ecs_ram_role (#60649)
- Support schema change for complex types in Iceberg external tables (#60169)
- Support score range filter pushdown (min_score semantics) (#60997)
- Support multiple tokenize index in one column (#60415)
- Add x509 cert based auth framework (#60098)
- Support dfs.client.use.datanode.hostname for libhdfs3 (#59915)
Improvements
- Improve password validation to align with MySQL STRONG policy (#60188)
- Optimize the performance of parquet dict decoder (#59681)
- Supports dynamically changing the rate limiter config (#59465)
- Split metadata scan ranges per split (#60257)
- Optimize Iceberg rewrite_data_files to avoid generating excessive small files (#60063)
- Use COUNT_ON_INDEX on variant subcolumns (#60404)
- Optimize file split size for multi-catalog (#60637)
- Limit the download rate of warmup task (#60180)
- Optimize performance of certain time field functions used with FROM_UNIXTIME (#60843)
- Add fold constant for PREVIOUS_DAY (#60755)
- Support PostgreSQL partition table sync for streaming job (#60560)
Bugfixes
- Fix crash caused by concurrent compaction accessing shared sample_infos (#60376)
- Fix unnecessary conflict range in lazy commit (#60274)
- Fix routine load task schedule stuck after create task fail (#60143)
- Fix DELETE/UPDATE cannot resolve column when table alias uses AS keyword (#60335)
- Fix boolean query AllScorer combination handling (#60438)
- Fix query cache not hit when use sort and one phase aggregation (#60298)
- Modify ORC reader to make error reporting more accurate (#60234)
- Fix prune nested column maybe throw NullPointerException (#60395)
- Fix partitions function throw error when query internal table without specifying catalog condition (#60247)
- Fix backends udf return rows incompatible with show backends command (#60210)
- Fix AggregateNode compute digest of query cache should consider sortByGroupKey (#60431)
- Fix predicate should be pushed down when conjunct contains CAST for Variant (#60485)
- Treat no such key as empty response when listing objects (part 2) for S3 (#60286)
- Fix empty string MATCH on keyword index returning wrong results (#60500)
- Fix PhysicalDictionarySink.resetLogicalProperties() does not reset logical properties properly (#60495)
- Fix %f (microseconds) format specifier in str_to_date (#60632)
- Throw exception for date_floor/ceil instead of return NULL (#60633)
- Fix S3-compatible storage must support temporary credentials when vendor credentials are configured at catalog level (#60232)
- Fix empty InsertIntoDictionaryCommand originSql (#60631)
- Fix width_bucket did not enforce that the fourth argument must be a constant (#60643)
- Validate mode parameter in search() DSL options (#60785)
- Fix range search prepare failed on NULL literal for ANN (#60564)
- Make dot match newline in regexp_fn by default (#60831)
- Fix Azure Storage Vault endpoint always using HTTP instead of HTTPS (#60854)
- Add rest external catalog for gson compatibility for Paimon (#60917)
- Disable strict mode for variant internal cast to fix INSERT INTO SELECT returning all NULLs (#60900)
- Default to the Default chain when S3 role_arn is set without provider_type (#60822)
- Fix parquet reader lazy materialization cannot filter (#60474)
- Fix point query ignoring session timezone for functions like from_unixtime (#60913)
- Make AND/OR/NOT operators case-sensitive in search DSL (#59747)
- Decouple min pipeline executor size from ConnectContext (#60958)
- Fix Azure endpoint detection for sovereign clouds with force-global config support (#60903)
- Backport search() function improvements and bug fixes (#61028)
- Fix encrypt key is not case sensitive (#60288)
- Make hive compress split assertion BE-aware (#60947)
- Support CLIENT_DEPRECATE_EOF to fix empty result with MySQL driver 9.5.0 (#61050)
- Fix basic and icu tokenizer can not be customized in custom analyzer (#60506)
