Release 0.144

    Querying bucketed tables in the Hive connector may produce incorrect results. This is fixed in Release 0.144.1, and .

    • Fix already exists check when adding a column to be case-insensitive.

    • Fix correctness issue when complex grouping operations have a partitioned source.

    • Fix potential memory starvation when a query is run with .

    • Queries run with resource_overcommit=true may now be killed before they reach query.max-memory if the cluster is low on memory.

    • Discard output stage JSON from completion event when it is very long. This limit can be configured with .

    • Improve performance when processing results in CLI and JDBC driver.

    • Improve performance of queries.

    Hive Changes

    • Fix ORC reader to actually use hive.orc.stream-buffer-size configuration property.