Release 374 (17 Mar 2022)#

General#

  • Add support for query parameters in CREATE SCHEMA. (#11485)

  • Improve performance when reading from S3-based spool for fault-tolerant execution. (#11050)

  • Improve performance of queries with GROUP BY clauses. (#11392)

  • Improve performance of GROUP BY with a large number of groups. (#11011)

  • Improve handling of queries where individual tasks require lots of memory when retry-policy is set to TASK. (#10432)

  • Produce better query plans by improving cost-based-optimizer estimates in the presence of correlated columns. (#11324)

  • Fix memory accounting and improve performance for queries involving certain variable-width data types such as varchar or varbinary. (#11315)

  • Fix performance regression for GROUP BY queries. (#11234)

  • Fix trim, ltrim and rtim function results when the argument is char type. Previously, it returned padded results as char type. It returns varchar type without padding now. (#11440)

JDBC driver#

  • Add support for DatabaseMetaData.getImportedKeys. (#8708)

  • Fix Driver.getPropertyInfo(), and validate allowed properties. (#10624)

CLI#

  • Add support for selecting Vim or Emacs editing modes with the --editing-mode command line argument. (#3377)

Cassandra connector#

ClickHouse connector#

  • Add support for uint8, uint16, uint32 and uint64 types. (#11490)

Delta Lake connector#

  • Allow specifying STS endpoint to be used when connecting to S3. (#10169)

  • Fix query failures due to exhausted file system resources after DELETE or UPDATE. (#11418)

Hive connector#

  • Allow specifying STS endpoint to be used when connecting to S3. (#10169)

  • Fix shared metadata caching with Hive ACID tables. (#11443)

Iceberg connector#

  • Allow specifying STS endpoint to be used when connecting to S3. (#10169)

  • Add support for using Glue metastore as Iceberg catalog. (#10845)

MongoDB connector#

PostgreSQL connector#

  • Improve performance of queries involving OR with simple comparisons and LIKE predicates by pushing predicate computation to the PostgreSQL database. (#11086)

  • Improve performance of aggregation queries with certain complex predicates by computing predicates and aggregations within PostgreSQL. (#11083)

  • Fix possible connection leak when connecting to PostgreSQL failed. (#11449)

SingleStore (MemSQL) connector#

  • The connector now uses the official Single Store JDBC Driver. As a result, connection-url in catalog configuration files needs to be updated from jdbc:mariadb:... to jdbc:singlestore:.... (#10669)

  • Deprecate memsql as the connector name. We recommend using singlestore in the connector.name configuration property. (#11459)