Release 373 (9 Mar 2022)#
General#
Add Delta Lake коннектор. (#11296, #10897)
Improve query performance by reducing overhead of cluster internal communication. (#11146)
Handle
varchar
totimestamp
conversion errors intry()
. (#11259)Add redirection awareness for
DROP COLUMN
task. (#11304)Add redirection awareness for
RENAME COLUMN
task. (#11226)Disallow table redirections in
SHOW GRANTS
statement. (#11270)Allow low memory killer to abort individual tasks when
retry-mode
is set toTASK
. This requiresquery.low-memory-killer.policy
set tototal-reservation-on-blocked-nodes
. (#11129)Fix incorrect results when distinct or ordered aggregation are used and spilling is enabled. (#11353)
Web UI#
Add CPU time, scheduled time, and cumulative memory statistics regarding failed tasks in a query. (#10754)
BigQuery connector#
Allow configuring view expiration time via the
bigquery.view-expire-duration
config property. (#11272)
Elasticsearch connector#
Improve performance of queries involving
LIKE
by pushing predicate computation to the Elasticsearch cluster. (#7994, #11308)
Hive connector#
Support access to S3 via a HTTP proxy. (#11255)
Improve query performance by better estimating partitioned tables statistics. (#11333)
Prevent failure for queries with the final number of partitions below
HIVE_EXCEEDED_PARTITION_LIMIT
. (#10215)Fix issue where duplicate rows could be inserted into a partition when
insert_existing_partitions_behavior
was set toOVERWRITE
andretry-policy
wasTASK
. (#11196)Fix failure when querying Hive views containing column aliases that differ in case only. (#11159)
Iceberg connector#
Support access to S3 via a HTTP proxy. (#11255)
Delete table data when dropping table. (#11062)
Fix
SHOW TABLES
failure when a materialized view is removed during query execution. (#10976)Fix query failure when reading from
information_schema.tables
orinformation_schema.columns
and a materialized view is removed during query execution. (#10976)
Oracle connector#
Fix query failure when performing concurrent write operations. (#11318)
Phoenix connector#
Prevent writing incorrect results when arrays contain
null
values. (#11351)
PostgreSQL connector#
Improve performance of queries involving
LIKE
by pushing predicate computation to the underlying database. (#11045)
SQL Server connector#
Fix incorrect results when querying SQL Server
tinyint
columns by mapping them to Trinosmallint
. (#11209)
SPI#
Add CPU time, scheduled time, and cumulative memory statistics regarding failed tasks in a query to query-completion events. (#10734)