Release 412 (5 Apr 2023)#
General#
Add support for aggregate functions and parameters as arguments for the
json_object()
andjson_array()
functions. (#16489, #16523, #16525)Expose optimizer rule execution statistics in query statistics. The number of rules for which statistics are collected can be limited with the
query.reported-rule-stats-limit
configuration property. (#2578)Add the
exclude_columns()
table function. (#16584)Allow disabling the use of the cost-based optimizer to determine partitioning of a stage with the
optimizer.use-cost-based-partitioning
configuration property or theuse_cost_based_partitioning
session property. (#16781)Improve performance of queries involving table functions with table arguments. (#16012)
Improve latency for small queries when fault-tolerant execution is enabled. (#16103)
Fix failure when querying a nested field of a
row
type in queries involvingORDER BY ... LIMIT
. (#16768)
JDBC driver#
Allow configuring a custom DNS resolver. (#16647)
ClickHouse connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to ClickHouse. (#16452)
Delta Lake connector#
Hive connector#
Iceberg connector#
Fix incorrect results for the
migrate
procedure when the table location contains a hidden directory. (#16779)
Ignite connector#
MariaDB connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to MariaDB. (#16452)
MySQL connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to MySQL. (#16452)
Oracle connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to Oracle. (#16452)
PostgreSQL connector#
Add support for table comments. (#16135)
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to PostgreSQL. (#16452)
Redshift connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to Redshift. (#16452)
SQL Server connector#
Improve performance of queries involving
sum(DISTINCT ...)
by pushing computation down to SQL Server. (#16452)
SPI#
Allow table functions to return anonymous columns. (#16584)