Release 396 (15 Sep 2022)#
General#
Improve performance of queries that process string data. (#12798)
Fix failure when querying views that use table functions. (#13944)
BigQuery connector#
Add support for column and table comments. (#13882)
ClickHouse connector#
Improve performance when pushing down large lists of predicates by raising the default threshold before the predicate is compacted. (#14029)
Fix mapping to the ClickHouse
Date
andDateTime
types to include the full range of possible values. (#11116)Fix failure when specifying a table or column comment that contains special characters. (#14058)
Delta Lake connector#
Add support for writing to tables using version 3 of the writer. This does not yet include support for
CHECK
constraints. (#14068)Add support for reading tables with the table property
delta.columnMapping.mode
set toid
. (#13629)Improve performance when writing structural data types to Parquet files. (#13714)
Attempt to undo the operation when writing a checkpoint file fails. (#14108)
Fix performance regression when reading Parquet data. (#14094)
Hive connector#
Add verification for files written by the optimized Parquet writer. This can be configured with the
parquet.optimized-writer.validation-percentage
configuration property or theparquet_optimized_writer_validation_percentage
session property. (#13246)Improve optimized Parquet writer performance for structural data types. (#13714)
Fix performance regression in reading Parquet files. (#14094)
Iceberg connector#
Improve performance when writing structural data types to Parquet files. (#13714)
Improve performance of queries that contain predicates involving
date_trunc
ondate
,timestamp
ortimestamp with time zone
partition columns. (#14011)Fix incorrect results from using the
[VERSION | TIMESTAMP] AS OF
clause when the snapshot’s schema differs from the current schema of the table. (#14064)Prevent
No bucket node map
failures when inserting data. (#13960)Fix performance regression when reading Parquet data introduced in Trino version 394. (#14094)
MariaDB connector#
Fix failure when using special characters in a table or column comment when creating a table. (#14058)
MySQL connector#
Fix failure when using special characters in a table or column comment when creating a table. (#14058)
Oracle connector#
Fix failure when setting a column comment with special characters. (#14058)
Phoenix connector#
Improve performance when pushing down large lists of predicates by raising the default threshold before the predicate is compacted. (#14029)
PostgreSQL connector#
Fix failure when setting a column comment with special characters. (#14058)
Redshift connector#
Fix failure when setting a column comment with special characters. (#14058)
SPI#
Add the
SystemAccessControl.checkCanGrantExecuteFunctionPrivilege
overload, which needs to be implemented to allow views that use table functions. (#13944)Add the
ConnectorMetadata.applyJoin
overload. It provides the connector with a join condition that is as complete as possible to represent usingConnectorExpression
. Deprecate the previous version ofConnectorMetadata.applyJoin
. (#13943)