12.55. Release 0.181

General Changes

  • Fix query failure and memory usage tracking when query containstransform_keys() or transform_values().
  • Prevent CREATE TABLE IF NOT EXISTS queries from ever failing with “Table already exists”.
  • Fix query failure when ORDER BY expressions reference columns that are used inthe GROUP BY clause by their fully-qualified name.
  • Fix excessive GC overhead caused by large arrays and maps containing VARCHAR elements.
  • Improve error handling when passing too many arguments to variousfunctions or operators that take a variable number of arguments.
  • Improve performance of count(*) aggregations over subqueries with knownconstant cardinality.
  • Add VERBOSE option for EXPLAIN ANALYZE that provides additionallow-level details about query performance.
  • Add per-task distribution information to the output of EXPLAIN ANALYZE.
  • Add support for DROP COLUMN in ALTER TABLE.
  • Change local scheduler to prevent starvation of long running querieswhen the cluster is under constant load from short queries. The newbehavior is disabled by default and can be enabled by setting theconfig property task.level-absolute-priority=true.
  • Improve the fairness of the local scheduler such that long-running querieswhich spend more time on the CPU per scheduling quanta (e.g., due toslow connectors) do not get a disproportionate share of CPU. The newbehavior is disabled by default and can be enabled by setting theconfig property task.legacy-scheduling-behavior=false.
  • Add a config option to control the prioritization of queries based onelapsed scheduled time. The task.level-time-multiplier propertycontrols the target scheduled time of a level relative to the nextlevel. Higher values for this property increase the fraction of CPUthat will be allocated to shorter queries. This config property onlyhas an effect when task.level-absolute-priority=true andtask.legacy-scheduling-behavior=false.

Hive Changes

  • Fix potential native memory leak when writing tables using RCFile.
  • Correctly categorize certain errors when writing tables using RCFile.
  • Decrease the number of file system metadata calls when reading tables.
  • Add support for dropping columns.

JDBC Driver Changes

  • Add support for query cancellation using Statement.cancel().

PostgreSQL Changes

  • Add support for operations on external tables.

Accumulo Changes

  • Improve query performance by scanning index ranges in parallel.

SPI Changes

  • Fix regression that broke serialization for SchemaTableName.
  • Add access control check for DROP COLUMN.