Release 408 (23 Feb 2023)#

General#

  • Add physical input read time to query statistics and the output of EXPLAIN ANALYZE. (#16190)

  • Fix query failure for queries involving joins or aggregations with a structural type that contains NULL elements. (#16140)

Security#

  • Deprecate using groups with OAuth 2.0 authentication, and rename the http-server.authentication.oauth2.groups-field configuration property to deprecated.http-server.authentication.oauth2.groups-field. (#15669)

CLI#

  • Add AUTO output format which switches from ALIGNED to VERTICAL if the output doesn’t fit the current terminal. (#12208)

  • Add --pager and --history-file options to match the existing TRINO_PAGER and TRINO_HISTORY_FILE environmental variables. Also allow setting these options in a configuration file. (#16151)

BigQuery connector#

  • Add support for writing decimal types to BigQuery. (#16145)

Delta Lake connector#

  • Rename the connector to delta_lake. The old name delta-lake is now deprecated and will be removed in a future release. (#13931)

  • Add support for creating tables with the Trino change_data_feed_enabled table property. (#16129)

  • Improve query performance on tables that Trino has written to with INSERT. (#16026)

  • Improve performance of reading structural types from Parquet files. This optimization can be disabled with the parquet_optimized_nested_reader_enabled catalog session property or the parquet.optimized-nested-reader.enabled catalog configuration property. (#16177)

  • Retry dropping Delta tables registered in the Glue catalog to avoid failures due to concurrent modifications. (#13199)

  • Allow updating the reader_version and writer_version table properties. (#15932)

  • Fix inaccurate change data feed entries for MERGE queries. (#16127)

  • Fix performance regression when writing to partitioned tables if table statistics are absent. (#16152)

Hive connector#

  • Remove support for the deprecated hive-hadoop2 connector name, requiring the connector.name property to be set to hive. (#16166)

  • Retry dropping Delta tables registered in the Glue catalog to avoid failures due to concurrent modifications. (#13199)

  • Fix performance regression when writing to partitioned tables if table statistics are absent. (#16152)

Iceberg connector#

  • Reduce memory usage when reading $files system tables. (#15991)

  • Require the iceberg.jdbc-catalog.driver-class configuration property to be set to prevent a «driver not found» error after initialization. (#16196)

  • Fix performance regression when writing to partitioned tables if table statistics are absent. (#16152)

Ignite connector#

SingleStore connector#

  • Remove support for the deprecated memsql connector name, requiring the connector.name property to be set to singlestore. (#16180)

SQL Server connector#

  • Add support for pushing down =, <> and IN predicates over text columns if the column uses a case-sensitive collation within SQL Server. (#15714)

Thrift connector#

  • Rename the connector to trino_thrift. The old name trino-thrift is now deprecated and will be removed in a future release. (#13931)