Skip to main content

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 and GROUP BY operations. (#25127)
  • Improve performance of GROUP BY and DISTINCT aggregations when spilling to disk is enabled
    or grouping by row, array, or map columns (#25294)
  • Fix failure when setting comments on columns with upper case letters. (#25297)
  • Fix potential query failure when retry_policy set to TASK (#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​

Phoenix connector​

SPI​

  • Add SourcePage interface and ConnectorPageSource.getNextSourcePage(). (#24011)
  • Deprecate ConnectorPageSource.getNextPage() for removal. (#24011)