Release 0.222

Warning

This release contains a regression which may cause Presto to incorrectly remove certain CAST operations.

General Changes

Fix incorrect results when dividing certain decimal numbers (#12930).
Fix planning failure for certain join queries caused by invalid distribution type (#12354).
Add support for automatically retrying failed buckets when using grouped execution. Currently this is supported for certain INSERT queries using the Hive connector. This can be enabled with the recoverable_grouped_execution session property or the recoverable-grouped-execution-enabled configuration property (#12124).
Add support for grouped execution for queries with no joins or aggregations. This can be enabled with the grouped_execution_for_eligible_table_scans session property or the experimental.grouped-execution-for-eligible-table-scans-enabled configuration property (#12934).
Add configuration property max-concurrent-materializations and session property max_concurrent_materializations to limit the number of plan sections that will run concurrently when using materialized exchanges.
Add support for computing approx_distinct() over BingTile values.
Add merge_hll() to merge an array of HyperLogLogs.
Add bitwise shift operations, bitwise_arithmetic_shift_right(), bitwise_logical_shift_right() and bitwise_shift_left().

Fix failures for information_schema queries when a table has an invalid storage format.
Improve query execution time over bucketed table with large buckets.
Add config property hive.metastore.glue.catalogid to configure the Glue catalog ID.

Add experimental interface ConnectorPlanOptimizer to allow connectors to participate in query plan optimization (e.g., filter pushdown) (#13102).
Rename LogicalRowExpressions::TRUE and LogicalRowExpressions::FALSE to LogicalRowExpressions::TRUE_CONSTANT and LogicalRowExpressions::FALSE_CONSTANT respectively to avoid collision with java.lang.Boolean.