Release 450 (19 Jun 2024)#

General#

Add support for specifying an Azure blob endpoint for accessing spooling in fault-tolerant execution with the exchange.azure.endpoint configuration property. (#22218)
Expose driver execution statistics via JMX. (#22427)
Improve performance of the first_value() and last_value() functions. (#22092)
Improve performance for large clusters under heavy workloads. (#22039)
Improve performance of queries with simple predicates. This optimization can be disabled using the experimental.columnar-filter-evaluation.enabled configuration property or the columnar_filter_evaluation_enabled session property. (#21375)
⚠️ Improve performance of aggregations containing a DISTINCT clause, and replace the optimizer.mark-distinct-strategy and optimizer.optimize-mixed-distinct-aggregations configuration properties with the new optimizer.distinct-aggregations-strategy property. (#21907)
Improve performance of reading JSON files. (#22348)
Improve performance for the date_trunc(), date_add(), and date_diff() functions. (#22192)
Fix failure when loading the OpenLineage event listener. (#22228)
Fix potential incorrect results when metadata or table data in certain connectors is updated or deleted. (#22285)

Add support for using web identity exclusively for authentication when running on Amazon EKS with the legacy S3 file system enabled. This can be configured via the trino.s3.use-web-identity-token-credentials-provider property. (#22162)
Add support for exclusively using web identity for authentication when using Amazon EKS with IAM roles by setting the s3.use-web-identity-token-credentials-provider configuration property. (#22163)

Add support for the assumeNullCatalogMeansCurrent connection property. When enabled, a null value for the catalog parameter in DatabaseMetaData methods is assumed to mean the current catalog. If no current catalog is set, the behaviour is unmodified. (#20866)

Add support for metadata caching when the bigquery.case-insensitive-name-matching configuration property is enabled. (#10740)
⚠️ Automatically configure BigQuery scan parallelism, and remove the bigquery.parallelism configuration property. (#22279)

Fix incorrect results when specifying a value for the cassandra.partition-size-for-batch-select configuration property. (#21940)

Improve performance of ORDER BY ... LIMIT on non-textual types by pushing execution down to the underlying database. (#22174)

Add support for concurrent UPDATE, MERGE, and DELETE queries. (#21727)
Add support for using table statistics with TIMESTAMP types. (#21878)
Add support for reading tables with type widening. (#21756)
Set the default value for the s3.max-connections configuration property to 500. (#22209)
Fix failure when reading a TIMESTAMP value after the year 9999. (#22184)
Fix failure when reading tables with the unsupported variant type. (#22310)
Add support for reading UniForm tables. (#22106)

Add support for changing a column’s type from integer to varchar and decimal to varchar, respectively, in unpartitioned tables. (#22246, #22293)
Add support for changing a column’s type from double to varchar in unpartitioned tables using Parquet files. (#22277)
Add support for changing a column’s type from float to varchar. (#22291)
Set the default value for the s3.max-connections configuration property to 500. (#22209)

Set the default value for the s3.max-connections configuration property to 500. (#22209)

Add support for the TRUNCATE statement. (#22340)
⚠️ Add support for V2 of the Nessie REST API. Previous behavior can be restored by setting the iceberg.nessie-catalog.client-api-version configuration property to V1. (#22215)
Improve performance when reading by populating split_offsets in file metadata. (#9018)
Set the default value for the s3.max-connections configuration property to 500. (#22209)
Fix failure when reading Parquet files that don’t have field-id on structured types. (#22347)