Release 446 (1 May 2024)#

General#

  • Improve performance of INSERT statements into partitioned tables when the retry_policy configuration property is set to TASK. (#21661)

  • Improve performance of queries with complex grouping operations. (#21726)

  • Reduce delay before killing queries when the cluster runs out of memory. (#21719)

  • Prevent assigning null values to non-null columns as part of a MERGE statement. (#21619)

  • Fix CREATE CATALOG statements including quotes in catalog names. (#21399)

  • Fix potential query failure when a column name ends with a :. (#21676)

  • Fix potential query failure when a SQL user-defined functions contains a label reference in a LEAVE, ITERATE, REPEAT, or WHILE statement. (#21682)

  • Fix query failure when SQL user-defined functions use the NULLIF or BETWEEN functions. (#19820)

  • Fix potential query failure due to worker nodes running out of memory in concurrent scenarios. (#21706)

BigQuery connector#

  • Improve performance when listing table comments. (#21581)

  • ⚠️ Breaking change: Enable bigquery.arrow-serialization.enabled by default. This requires --add-opens=java.base/java.nio=ALL-UNNAMED in jvm-config. (#21580)

Delta Lake connector#

  • Fix failure when reading from Azure file storage and the schema, table, or column name contains non-alphanumeric characters. (#21586)

  • Fix incorrect results when reading a partitioned table with a deletion vector. (#21737)

Hive connector#

  • Add support for reading S3 objects restored from Glacier storage. (#21164)

  • Fix failure when reading from Azure file storage and the schema, table, or column name contains non-alphanumeric characters. (#21586)

  • Fix failure when listing Hive views with unsupported syntax. (#21748)

Iceberg connector#

  • Add support for the Snowflake catalog. (#19362)

  • Automatically use varchar as a type during table creation when char is specified. (#19336, #21515)

  • Deprecate the schema and table arguments for the table_changes function in favor of schema_name and table_name, respectively. (#21698)

  • Fix failure when executing the migrate procedure with partitioned Hive tables on Glue. (#21391)

  • Fix failure when reading from Azure file storage and the schema, table, or column name contains non-alphanumeric characters. (#21586)

Pinot connector#

  • Fix query failure when a predicate contains a '. (#21681)

Snowflake connector#

  • Add support for the unsupported-type-handling and jdbc-types-mapped-to-varchar type mapping configuration properties. (#21528)

SPI#

  • Remove support for @RemoveInput as an annotation for aggregation functions. A WindowAggregation can be declared in @AggregationFunction instead, which supports input removal. (#21349)

  • Extend QueryCompletionEvent with various aggregated, per-stage, per-task distribution statistics. New information is available in QueryCompletedEvent.statistics.taskStatistics. (#21694)