Presto 0.242 Documentation

13.2. Release 0.241

13.2. Release 0.241

General Changes

  • Fix incorrect results from function classification_precision() 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 experimental.enable-dynamic-filtering configuration property and enable_dynamic_filtering system session property. For more information please refer to (#15077).
  • Add new warning message for UNION queries without ALL or DISTINCT keywords.
  • Downgrade the ZSTD JNI compressor version to resolve the frequent excessive GC events introduced in version 0.238.

Security Changes

  • Implement REST endpoint authorization in Presto. See Authorization.

JDBC Changes

  • Add support for ResultSet.getStatement.
  • Add support for Oracle JDBC connections.

Hive Changes

  • Fix several memory accounting bugs in OrcRecordReader and StreamReader.
  • Add procedure system.sync_partition_metadata() to synchronize the partitions in the metastore with the partitions that are physically in the file system.
  • Add support for direct recursive file listings in PrestoS3FileSystem.
  • 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 and parquet_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.

Verifier Changes

  • Fix an issue during determinism analysis that queries with LIMIT 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 and OperatorStatistics to QueryCompletedEvent, and remove stage and operator statistics from QueryStatistics.

Geospatial Changes