Closed martint closed 1 year ago
# General
* Fix failures for queries with remote partitioning exchange on a complex type (ROW, ARRAY, MAP) with values containing NULL elements. ({issue}`https://github.com/trinodb/trino/issues/16140`)
# Iceberg
* Reduce memory usage when reading `$files` system tables. ({issue}`15991`)
# Delta Lake
* Fix change data feed entries for `MERGE` queries. ({issue}`16127`)
# BigQuery
* Add support for writing BigQuery `numeric` and `bignumeric` types. ({issue}`16145`)
# SingleStore
* Remove support for deprecated `memsql` connector name. ({issue}`16180`)
# Hive, Hudi, Delta, Iceberg
* Fix performance regression for partitioned table writes when table statistics are absent. ({issue}`16152`)
# Delta Lake
* Support creating a table with `change_data_feed_enabled` table property. ({issue}`16129`)
# Iceberg
* Require `iceberg.jdbc-catalog.driver-class` config property to prevent driver not found error after the initialization. ({issue}`16196`)
# Hive, Delta
* Retry dropping delta tables registered on AWS Glue in case of failures due to concurrent modifications. ({issue}`13199`)
# Hive, Hudi, Delta, Iceberg
* Improve performance of reading ARRAY, MAP and ROW types from parquet files. This optimization is enabled by default. It can be disabled using the catalog session property `parquet_optimized_nested_reader_enabled` or the catalog configuration property `parquet.optimized-nested-reader.enabled`. ({issue}`16177`)
# SQL Server
* Add support for pushing down `=`, `!=` and `IN` predicates over text columns to SQL Server if the column uses a case-sensitive collation within SQL Server. ({issue}`15714`)
# General
* Report physical input read time in eventlistener QueryStatistics and EXPLAIN ANALYZE. ({issue}`16190`)
# Delta Lake
* Allow updating `reader_version` and `writer_version` table properties. ({issue}`15932`)
16026