Release 0.226#

General Changes#

  • Fix ST_Buffer for tiny geometries (#13194)

  • Fix ST_Centroid for tiny geometries (#10629)

  • Fix failure when reading unpartitioned table from some connectors with session property grouped_execution_for_eligible_table_scans turned on.

  • Improve memory tracking for geospatial query indexing and prevent worker node from crashing due to out of memory.

  • Improve memory usage of spatial joins.

  • Add peak running task count to statistics field of QueryCompletedEvent.

Hive Changes#

  • Fix computation for number of buckets accessed for max_buckets_for_grouped_execution.

  • Fix a bug where the bucket column was not available if max_buckets_for_grouped_execution was exceeded.

  • Fix high CPU usage when writing ORC files with too many row groups.

  • Improve parallelism of writes to hive bucketed tables by respecting the task.writer-count configuration property.

  • Add debug mode that allows reading from offline table and partition. This is controlled by session property offline_data_debug_mode_enabled.