Release 0.241
There is a bug in LambdaDefinitionExpression canonicalization introduced since 0.238. For more details, go to #15424.
Fix incorrect results from function introduced in release 0.239.
Improve performance for queries with broadcast or collocated joins by adding dynamic filtering and bucket pruning support. This can be enabled with the configuration property and
enable_dynamic_filtering
system session property. For more information please refer to (#15077).Add new warning message for
UNION
queries withoutALL
orDISTINCT
keywords.
Security Changes
- Implement REST endpoint authorization in Presto. See Authorization.
JDBC Changes
Add support for
ResultSet.getStatement
.Add support for Oracle JDBC connections.
Fix several memory accounting bugs in
OrcRecordReader
and .Add support for non-Hive types to Hive views. This support had been removed in 0.233. If a view uses an unsupported type for any columns, only a single dummy column will be saved in the metastore.
Add support for pushing dereferences into Parquet table scan, so that only the required nested column is read when other projected nested columns are in the same base column. This can be enabled with the
hive.enable-parquet-dereference-pushdown
Hive configuration property andparquet_batch_reader_verification_enabled
Hive session property.
Druid Changes
- Add support for data ingestion. Three modes are supported:
INSERT INTO SELECT
,CREATE TABLE AS
, and from local or HDFS folders.
Oracle Changes
- Add Oracle connector.
Fix an issue during determinism analysis that queries with clause are not identified as non-deterministic when a rerun of the control query fails.
Add support to allow multiple control and test clusters.
SPI Changes
- Add
StageStatistics
andOperatorStatistics
toQueryCompletedEvent
, and remove stage and operator statistics fromQueryStatistics
.
Geospatial Changes
Add
geometry_nearest_points()
to find nearest points of a pair of geometries.