Release 473 (19 Mar 2025)#
Warning
This release is broken and should not be used. GROUP BY and DISTINCT queries containing more than 33M unique groups can produce incorrect results. See: (#25381)
General#
Add support for array literals. (#25301)
Reduce the amount of memory required for
DISTINCT
andGROUP BY
operations. (#25127)Improve performance of
GROUP BY
andDISTINCT
aggregations when spilling to disk is enabled
or grouping byrow
,array
, ormap
columns (#25294)Fix failure when setting comments on columns with upper case letters. (#25297)
Fix potential query failure when
retry_policy
set toTASK
(#25217)
Security#
Add LDAP-based group provider. (#23900)
Fix column masks not being applied on view columns with upper case. (#24054)
BigQuery connector#
Fix failure when initializing the connector on a machine with more than 32 CPU cores. (#25228)
Delta Lake connector#
Remove the deprecated
glue-v1
metastore type. (#25201)Remove deprecated Databricks Unity catalog integration. (#25250)
Fix Glue endpoint URL override. (#25324)
Hive connector#
Remove the deprecated
glue-v1
metastore type. (#25201)Remove deprecated Databricks Unity catalog integration. (#25250)
Fix Glue endpoint URL override. (#25324)
Hudi connector#
Fix queries getting stuck when reading empty partitions. (#19506)
Remove the deprecated
glue-v1
metastore type. (#25201)Fix Glue endpoint URL override. (#25324)
Iceberg connector#
Set the
write.<filetype>.compression-codec
table property when creating new tables. (#24851)Expose additional properties in
$properties
tables. (#24812)Fix Glue endpoint URL override. (#25324)
Kudu connector#
Remove the Kudu connector. (#24417)
Phoenix connector#
Remove the Phoenix connector. (#24135)