Add support for correlated queries involving a VALUES clause with a single
row. (#15989)
Reduce memory usage for large schemas. This behavior can be enabled with the
query.remote-task.enable-adaptive-request-size configuration property and
configured with the query.remote-task.max-request-size,
query.remote-task.request-size-headroom, and
query.remote-task.guaranteed-splits-per-task configuration properties or
their respective session properties. (#15721)
Improve concurrency when small, concurrent queries are run on a large cluster
by automatically determining how many nodes to use for distributed joins
and aggregations. This can be managed with the query.max-hash-partition-count
(renamed from query.hash-partition-count) and query.min-hash-partition-count
configuration properties. (#15489)
Improve memory usage accounting for queries with dynamic filters. (#16110)
Improve query performance when a predicate evaluates to a null value. (#15744)
Improve performance of queries with joins on the output of global
aggregations. (#15858)
Improve performance of selective queries, queries that read a small number of
columns, and queries that process tables with large Parquet row groups or ORC
stripes. (#15579)
Improve performance of queries with window functions. (#15994)
Return an exit code of 100 when Trino crashes during startup. (#16113)
Fix precision loss when converting time values with a precision higher than
three and time with time zone values with lower precision. (#15861)
Fix potential incorrect results due to a query reporting normal completion
instead of failing. (#15917)
Fix connection errors caused by a reusable connection being closed. (#16121)
Fix incorrect results for queries involving an equality predicate in a WHERE
clause that is equal to a term of a SELECT clause in one of the branches of
a JOIN. (#16101)
Improve performance of reading decimal types from Parquet files. (#15713)
Improve performance of reading numeric types from Parquet files. (#15850)
Improve performance of reading string types from Parquet files. (#15897, #15923)
Improve performance of reading timestamp and boolean types from Parquet files. (#15954)
Improve performance of predicate pushdown to partitioned columns in tables
with a high number of partitions. (#16113)
Reduce server errors in high-load scenarios. This can be enabled with the
hive.s3.connect-ttl configuration property. (#16005)
Allow setting the hive.max-partitions-per-scan configuration property to a
value lower than the value set in hive.max-partitions-for-eager-load. (#16111)
Fix query failure when reading Parquet files written by Apache Impala. (#15942)
Fix listing relations failure when a Glue table has no table type set. (#15909)
Add support for the unregister_table procedure. (#15784)
Add support for register_table procedures in the JDBC catalog. (#15853)
Add support for specifying a user and password when connecting to the JDBC
catalog via the iceberg.jdbc-catalog.connection-user and
iceberg.jdbc-catalog.connection-password configuration properties. (#16040)
Add support for compacting manifests asynchronously, which can be enabled by
setting the iceberg.merge_manifests_on_write session property to false. (#14822)