-
Notifications
You must be signed in to change notification settings - Fork 127
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Co-authored-by: zhangli20 <[email protected]>
- Loading branch information
Showing
4 changed files
with
26 additions
and
37 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,23 +1,12 @@ | ||
# blaze-v2.0.9 | ||
# blaze-v2.0.9.1 | ||
|
||
## Features | ||
* Upgrades datafusion/arrow dependency version to v36/v50. | ||
* Supports max/min aggregation with complex types. | ||
* Supports json_tuple. | ||
* Introduce sonic-rs for json parsing. | ||
* Add stage id in operator metrics. | ||
* Implements writing table with dynamic partitions (not tested in spark303/spark333). | ||
* Supports failing-back nondeterministic expressions. | ||
* Supports "$[].xxx" jsonpath syntax in get_json_object(). | ||
|
||
## Performance | ||
* Improves batch serialization format and reduce compressed size. | ||
* Implements radix-based k-way merging used in shuffling and aggregating. | ||
* Improves performance of on-heap spilling. | ||
* Improves performance of SortExec. | ||
* Improves performance of AggExec. | ||
* Improves performance of collect_set/collect_list. | ||
* Supports adaptive batch size in ParquetScan, improving vectorized reading performance. | ||
* Supports directly spill to disk file when on-heap memory is full. | ||
|
||
## Bugfix | ||
* Fix concat_ws with empty batches. | ||
* Fix spark333 RenameExec incorrect ordering expressions. | ||
* Fix incorrect join type mapping in BroadcastNestedLoopJoin. | ||
* Fix decimal dividing with zero. | ||
* Fix incorrect parquet rowgroup pruning with files containing deprecated min/max values. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters