Release 398 (28 Sep 2022)#

General#

  • Add Hudi connector. (#10228)

  • Add metrics for the execution time of filters and projections to EXPLAIN ANALYZE VERBOSE. (#14135)

  • Show local cost estimates when using EXPLAIN. (#14268)

  • Fix timeouts happening too early because of improper handling of the node-scheduler.allowed-no-matching-node-period configuration property. (#14256)

  • Fix query failure for MERGE queries when task_writer_count is greater than one. (#14306)

Accumulo connector#

  • Add support for column comments when creating a new table. (#14114)

  • Move column mapping and index information into the output of DESCRIBE instead of a comment. (#14095)

BigQuery connector#

  • Fix improper escaping of backslash and newline characters. (#14254)

  • Fix query failure when the predicate involves a varchar value with a backslash. (#14254)

ClickHouse connector#

  • Upgrade minimum required Clickhouse version to 21.8. (#14112)

Delta Lake connector#

  • Improve performance when reading Parquet files for queries with predicates. (#14247)

Elasticsearch connector#

  • Deprecate support for query pass-through using the special <index>$query:<es-query> dynamic tables in favor of the raw_query table function. Legacy behavior can be re-enabled with the elasticsearch.legacy-pass-through-query.enabled configuration property. (#14015)

Hive connector#

  • Add support for partitioned views when legacy mode for view translation is enabled. (#14028)

  • Extend the flush_metadata_cache procedure to be able to flush table-related caches instead of only partition-related caches. (#14219)

  • Improve performance when reading Parquet files for queries with predicates. (#14247)

Iceberg connector#

  • Improve performance when reading Parquet files for queries with predicates. (#14247)

  • Fix potential table corruption when changing a table before it is known if committing to the Glue metastore has failed or succeeded. (#14174)

Pinot connector#

  • Add support for the timestamp type. (#10199)

SPI#

  • Extend ConnectorMetadata.getStatisticsCollectionMetadata to allow the connector to request the computation of any aggregation function during stats collection. (#14233)