Release 447 (8 May 2024)

General

• Add support for SHOW CREATE FUNCTION. (#21809)

• Add support for the bitwise_xor_agg() aggregation function. (#21436)

• ⚠️ Breaking change: Require JDK 22 to run Trino. (#20980)

• Improve performance of ORDER BY queries with LIMIT on large data sets. (#21761)

• Improve performance of queries containing the rank() or row_number() window functions. (#21639)

• Improve performance of correlated queries with EXISTS. (#21422)

• Fix potential failure for expressions involving try_cast(parse_json(...)). (#21877)

CLI

• Fix incorrect error location markers for SQL routines causing CLI to print exceptions. (#21357)

Delta Lake connector

• Add support for concurrent DELETE and TRUNCATE queries. (#18521)

• Fix under-accounting of memory usage when writing strings to Parquet files. (#21745)

Hive connector

• Add support for metastore caching on tables that have not been analyzed, which can be enabled with the hive.metastore-cache.cache-missing-stats and hive.metastore-cache.cache-missing-partitions configuration properties. (#21822)

• Fix under-accounting of memory usage when writing strings to Parquet files. (#21745)

• Fix failure when translating Hive views that contain EXISTS clauses. (#21829)

Hudi connector

• Fix under-accounting of memory usage when writing strings to Parquet files. (#21745)

Iceberg connector

• Fix under-accounting of memory usage when writing strings to Parquet files. (#21745)

Phoenix connector

• ⚠️ Breaking change: Remove support for Phoenix versions 5.1.x and earlier. (#21569)

Pinot connector

• Add support for specifying an explicit broker URL with the pinot.broker-url configuration property. (#17791)

Redshift connector

• ⚠️ Breaking change: Remove deprecated legacy type mapping and the associated redshift.use-legacy-type-mapping configuration property. (#21855)