Release 372 (2 Mar 2022)#
General#
Add
trim_array()
function. (#11238)Improve handling of prepared statements with long query text by compressing them within HTTP headers. This can be configured or disabled using the
protocol.v1.prepared-statement-compression.length-threshold
configuration property. (#11098)Improve performance of specific queries which compare table columns of type
timestamp
withdate
literals. (#11170)Add redirection awareness for
ADD COLUMN
,DROP TABLE
,COMMENT
tasks. (#11072)Remove support for reserved memory pool. Configuration property
experimental.reserved-pool-disabled
can no longer be used. (#6677)Ensure memory is released completely after query completion. (#11030)
Fix certain queries failing due to dictionary compacting error. (#11080)
Fix
SET SESSION
andRESET SESSION
not working for catalogs which include special characters in their name. (#11171)Fix bug where queries were not transitioned to
RUNNING
state when task-level retries were enabled. (#11198)
Security#
Allow configuration of connect and read timeouts for LDAP authentication. (#10925)
Docker image#
Add a health check to the Docker container image. (#10413)
JDBC driver#
Base-JDBC connector library#
Fix spurious query failures when metadata cache is not enabled and data access depends on the session state. (#11068)
Accumulo connector#
Fix incorrect results when querying
date
type columns. (#11055)
Cassandra connector#
Fix incorrect results when filtering partition keys without projections. (#11001)
ClickHouse connector#
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
Druid connector#
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
Hive connector#
Allow redirects of Iceberg or Delta tables which have no
StorageDescriptor
in the Glue metastore. (#11092)Stop logging spurious failures when storage caching is used. (#11101)
Allow reading Zstandard compressed Avro files. (#11090)
Fix incorrect query results after writes to a table when directory caching is enabled enabled with the
hive.file-status-cache-tables
configuration property. (#10621)Fix potential query failures for queries writing data to tables backed by S3. (#11089)
Iceberg connector#
Add support for
COMMENT ON COLUMN
statement. (#11143)Improve query performance after table schema evolved, by fixing the connector to support table stats in such case. (#11091)
Fix potential query failures for queries writing data to tables backed by S3. (#11089)
Prevent query failure from dereference pushdown when a column has a comment. (#11104)
Kudu connector#
Add support for Kerberos authentication. (#10953)
MongoDB connector#
Map MongoDB
bindata
type to Trinovarbinary
type if explicit schema does not exist. (#11122)
MySQL connector#
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
Oracle connector#
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
Phoenix connector#
PostgreSQL connector#
Redshift connector#
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
SingleStore (MemSQL) connector#
Fix spurious query failures when metadata cache is not enabled, and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
SQL Server connector#
Update JDBC driver to 10.2.0. The new version automatically enables TLS and certificate verification. Update the TLS configuration to keep the old behavior. (#10898)
Fix spurious query failures when metadata cache is not enabled and extra credentials with
user-credential-name
orpassword-credential-name
are used to access data. (#11068)
SPI#
Pass more information about predicates in
ConnectorMetadata#applyFilter
invocation. The predicates that cannot be represented with aTupleDomain
are available viaConstraint.getExpression()
. (#7994)