Release 340 (8 Aug 2020)#
General#
Add support for query parameters in
LIMIT
,OFFSET
andFETCH FIRST
clauses. (#4529, #4601)Add experimental support for recursive queries. (#4250)
Add
bitwise_left_shift()
,bitwise_right_shift()
andbitwise_right_shift_arithmetic()
. (#740)Add
luhn_check()
. (#4011)Add
IF EXISTS
andIF NOT EXISTS
syntax toALTER TABLE
. (#4651)Include remote host in error info for page transport errors. (#4511)
Improve minimum latency for dynamic partition pruning. (#4388)
Reduce cluster load by cancelling query stages from which data is no longer required. (#4290)
Reduce query memory usage by improving retained size estimation for
VARCHAR
andCHAR
types. (#4123)Improve query performance for queries containing
starts_with()
. (#4669)Improve performance of queries that use
DECIMAL
data type. (#4730)Fix failure when
GROUP BY
clause contains duplicate expressions. (#4609)Fix potential hang during query planning (#4635).
Security#
Fix unprivileged access to table’s schema via
CREATE TABLE LIKE
. (#4472)
JDBC driver#
Elasticsearch connector#
Fix failure when index mapping is missing. (#4535)
Hive connector#
Allow creating a table with
external_location
when schema’s location is not valid. (#4069)Add read support for tables that were created as non-transactional and converted to be transactional later. (#2293)
Allow creation of transactional tables. Note that writing to transactional tables is not yet supported. (#4516)
Add
hive.metastore.glue.max-error-retries
configuration property for the number of retries performed when accessing the Glue metastore. (#4611)Support using Java KeyStore files for Thrift metastore TLS configuration. (#4432)
Expose hit rate statistics for Hive metastore cache via JMX. (#4458)
Improve performance when querying a table with large files and with
skip.header.line.count
property set to 1. (#4513)Improve performance of reading JSON tables. (#4705)
Fix query failure when S3 data location contains a
_$folder$
marker object. (#4552)Fix failure when referencing nested fields of a
ROW
type when table and partition metadata differs. (#3967)
Kafka connector#
MongoDB connector#
Add
mongodb.max-connection-idle-time
properties to limit the maximum idle time of a pooled connection. (#4483)