Release 306 (16 Mar 2019)#

General#

  • Fix planning failure for queries containing a LIMIT after a global aggregation. (#437)

  • Fix missing column types in EXPLAIN output. (#328)

  • Fix accounting of peak revocable memory reservation. (#413)

  • Fix double memory accounting for aggregations when spilling is active. (#413)

  • Fix excessive CPU usage that can occur when spilling for window functions. (#468)

  • Fix incorrect view name displayed by SHOW CREATE VIEW. (#433)

  • Allow specifying NOT NULL when creating tables or adding columns. (#418)

  • Add a config option (query.stage-count-warning-threshold) to specify a per-query threshold for the number of stages. When this threshold is exceeded, a TOO_MANY_STAGES warning is raised. (#330)

  • Support session property values with special characters (e.g., comma or equals sign). (#407)

  • Remove the deprecated.legacy-unnest-array-rows configuration option. The legacy behavior for UNNEST of arrays containing ROW values is no longer supported. (#430)

  • Remove the deprecated.legacy-row-field-ordinal-access configuration option. The legacy mechanism for accessing fields of anonymous ROW types is no longer supported. (#428)

  • Remove the deprecated.group-by-uses-equal configuration option. The legacy equality semantics for GROUP BY are not longer supported. (#432)

  • Remove the deprecated.legacy-map-subscript. The legacy behavior for the map subscript operator on missing keys is no longer supported. (#429)

  • Remove the deprecated.legacy-char-to-varchar-coercion configuration option. The legacy coercion rules between CHAR and VARCHAR types are no longer supported. (#431)

  • Remove deprecated distributed_join system property. Use join_distribution_type instead. (#452)

Hive connector#

  • Fix calling procedures immediately after startup, before any other queries are run. Previously, the procedure call would fail and also cause all subsequent Hive queries to fail. (#414)

  • Improve ORC reader performance for decoding REAL and DOUBLE types. (#465)

MySQL connector#

  • Allow creating or renaming tables, and adding, renaming, or dropping columns. (#418)

PostgreSQL connector#

  • Fix predicate pushdown for PostgreSQL ENUM type. (#408)

  • Allow creating or renaming tables, and adding, renaming, or dropping columns. (#418)

Redshift connector#

  • Allow creating or renaming tables, and adding, renaming, or dropping columns. (#418)

SQL Server connector#

  • Allow creating or renaming tables, and adding, renaming, or dropping columns. (#418)

Base-JDBC connector library#

  • Allow mapping column type to Presto type based on Block. (#454)

SPI#

  • Deprecate Table Layout APIs. Connectors can opt out of the legacy behavior by implementing ConnectorMetadata.usesLegacyTableLayouts(). (#420)

  • Add support for limit pushdown into connectors via the ConnectorMetadata.applyLimit() method. (#421)

  • Add time spent waiting for resources to QueryCompletedEvent. (#461)