Release 375 (28 Mar 2022)#

General#

  • Change behavior of ALTER TABLE qualified_name TO unqualified_name to leave the table in the schema. This is backwards-incompatible behavioral change. (#11282)

  • Disallow table redirections for the GRANT, REVOKE, and SET TABLE AUTHORIZATION table tasks. (#11302)

  • Improve performance of queries that contain filter predicates on aggregation results. (#11469)

  • Improve performance of queries that contain array_distinct(array_sort(…)) expressions. (#8777)

  • Fix SHOW CREATE TABLE to show actual table name in case of table redirections, so that the statement can be used to recreate the table. (#11604)

  • Fix scheduling for non-remotely accessible splits in fault-tolerant execution. (#11581)

  • Fix incorrect is_nullable value in information_schema.columns table. (#11613)

JDBC driver#

  • Allow PreparedStatement.close() to be called multiple times. (#11620)

  • Fix incorrect is_nullable value in DatabaseMetaData.getColumns() method. (#11613)

Cassandra connector#

  • Return 0 instead of row count in completed bytes. (#11644)

Delta Lake connector#

  • Add access control to drop_extended_stats and vacuum procedures. (#11633)

  • Fix incorrect query results when query executes concurrent with DROP TABLE. (#11562)

Hive connector#

  • Fix infinite loop in the query optimizer when query contains predicates on a struct field. (#11559)

  • Fix query failure when reading a partitioned table with a predicate on a partition column with a specific name such as table, key, order, and others. (#11512)

Iceberg connector#

  • Fix failure when query contains predicates on a struct field. (#11560)

  • Fix query failure when reading from $files system table after a table column has been dropped. (#11576)

Kudu connector#

  • Improve write performance by flushing operations in batches. (#11264)

  • Fix failure when multiple queries run concurrently and schema emulation is enabled. (#11264)

MongoDB connector#

  • Support predicate pushdown on boolean columns. (#11536)

  • Return 0 instead of row count in completed bytes. (#11679)

MySQL connector#

  • Add support for table comments. (#11211)

Pinot connector#

  • Fix handling of passthrough queries that contain aggregation functions. (#9137)

  • Fix incorrect results when aggregation functions on columns having non-lowercase names are pushed down to Pinot. (#9137, #10148)

  • Fix possible incorrect results when grouping on columns of array types. (#9781)

PostgreSQL connector#

  • Improve performance of queries involving OR with IS NULL, IS NOT NULL predicates, or involving NOT expression by pushing predicate computation to the PostgreSQL database. (#11514)

  • Improve performance of queries with predicates involving nullif function by pushing predicate computation to the PostgreSQL database. (#11532)

  • Improve performance of queries involving joins by pushing computation to the PostgreSQL database. (#11635)

  • Improve performance of queries involving predicates with arithmetic expressions by pushing predicate computation to the PostgreSQL database. (#11510)

  • Fix deletion of too much data when delete query involves a LIKE predicate. (#11615)

SPI#

  • Add processed input bytes and rows to query events in event listener. (#11623)

  • Remove deprecated constructors from ColumnMetadata. (#11621)