12.10. Release 320 (10 Oct 2019)
General Changes
- Fix incorrect parameter binding order for prepared statement execution when
parameters appear inside a
WITH
clause. (#1191)
- Fix planning failure for certain queries involving a mix of outer and
cross joins. (#1589)
- Improve performance of queries containing complex predicates. (#1515)
- Avoid unnecessary evaluation of redundant filters. (#1516)
- Improve performance of certain window functions when using bounded window
frames (e.g.,
ROWS BETWEEN ... PRECEDING AND ... FOLLOWING
). (#464)
- Add Kinesis Connector. (#476)
- Add
geometry_from_hadoop_shape()
. (#1593)
- Add
at_timezone()
. (#1612)
- Add
with_timezone()
. (#1612)
JDBC Driver Changes
- Only report warnings on
Statement
, not ResultSet
, as warnings
are not associated with reads of the ResultSet
. (#1640)
CLI Changes
- Add multi-line editing and syntax highlighting. (#1380)
Hive Connector Changes
- Add impersonation support for calls to the Hive metastore. This can be enabled using the
hive.metastore.thrift.impersonation.enabled
configuration property. (#43)
- Add caching support for Glue metastore. (#1625)
- Add separate configuration property
hive.hdfs.socks-proxy
for accessing HDFS via a
SOCKS proxy. Previously, it was controlled with the hive.metastore.thrift.client.socks-proxy
configuration property. (#1469)
MySQL Connector Changes
- Add
mysql.jdbc.use-information-schema
configuration property to control whether
the MySQL JDBC driver should use the MySQL information_schema
to answer metadata
queries. This may be helpful when diagnosing problems. (#1598)
PostgreSQL Connector Changes
- Add support for reading PostgreSQL system tables, e.g.,
pg_catalog
relations.
The functionality is disabled by default and can be enabled using the
postgresql.include-system-tables
configuration property. (#1527)
Elasticsearch Connector Changes
- Add support for
VARBINARY
, TIMESTAMP
, TINYINT
, SMALLINT
,
and REAL
data types. (#1639)
- Discover available tables and their schema dynamically. (#1639)
- Add support for special
_id
, _score
and _source
columns. (#1639)
- Add support for full text queries. (#1662)
SPI Changes
- Introduce a builder for
Identity
and deprecate its public constructors. (#1624)