Release 0.156#


Query may incorrectly produce NULL when no row qualifies for the aggregation if the optimize_mixed_distinct_aggregations session property or the optimizer.optimize-mixed-distinct-aggregations config option is enabled.


  • Fix potential correctness issue in queries that contain correlated scalar aggregation subqueries.

  • Fix query failure when using AT TIME ZONE in VALUES list.

  • Add support for quantified comparison predicates: ALL, ANY, and SOME.

  • Add support for ARRAY and ROW that contain NULL in checksum() aggregation.

  • Add support for filtered aggregations. Example: SELECT sum(a) FILTER (WHERE b > 0) FROM ...

  • Add a variant of from_unixtime() function that takes a timezone argument.

  • Improve performance of GROUP BY queries that compute a mix of distinct and non-distinct aggregations. This optimization can be turned on by setting the optimizer.optimize-mixed-distinct-aggregations configuration option or via the optimize_mixed_distinct_aggregations session property.

  • Change default task concurrency to 16.


  • Add support for legacy RCFile header version in new RCFile reader.


  • Support iso8601 data format for the hash row decoder.


  • Make ConnectorPageSink#finish() asynchronous.


These are backwards incompatible changes with the previous SPI. If you have written a plugin, you will need to update your code before deploying this release.