Release 0.270

Details

General Changes

  • Fix error classification when an invalid timezone is passed as a parameter to from_unixtime().

  • Improve performance of DISTINCT LIMIT N queries for N <= 10000. This can be enabled with the session property hash_based_distinct_limit_enabled or the configuration property hash-based-distinct-limit-enabled and the limit can be adjusted by using the session property hash_based_distinct_limit_threshold or the configuration property hash-based-distinct-limit-threshold.

  • Add last_day_of_month() UDF to return the last day of the month.

  • Add dynamic filtering support for right join.

  • Add support for any expression for dynamic filtering probe side.

  • Add new optimizer rule to simplify expressions like cardinality(map_keys(m)) into cardinality((m)). Same for map_values() function.

  • Add support for the following primitive types to Avro decoder: TINYINT, SMALLINT, INTEGER and REAL.

Hive Connector Changes

  • Remove the configuration property hive.parquet.fail-on-corrupted-statistics and the session property parquet_fail_with_corrupted_statistics.

Iceberg Connector Changes

  • Remove the configuration property iceberg.native-mode. Use iceberg.catalog.type instead.

Pinot Connector Changes

  • Add support for Pinot TIMESTAMP and JSON types.

  • Add support for Pinot version 0.9.3.

Credits

Amit Dutta, Arunachalam Thirupathi, Beinan, Chunxu Tang, James Petty, James Sun, Nikolay Laptev, Pedro Sereno, Pranjal Shankhdhar, Rongrong Zhong, Sreeni Viswanadha, Timothy Meehan, Xiang Fu, Yang Yang, Zhenxiao Luo, chuxiao, ericyuliu, wangjingdong, zhangyanbing