12.3. Release 327 (20 Dec 2019)
General Changes
- Fix join query failure when late materialization is enabled. (#2144)
- Fix failure of
word_stem()
for certain inputs. (#2145)
- Fix query failure when using
transform_values()
inside try()
and the transformation fails
for one of the rows. (#2315)
- Fix potential incorrect results for aggregations involving
FILTER (WHERE ...)
when the condition is a reference to a table column. (#2267)
- Allow renaming views with ALTER VIEW. (#1060)
- Add
error_type
and error_code
columns to system.runtime.queries
. (#2249)
- Rename
experimental.work-processor-pipelines
configuration property to experimental.late-materialization.enabled
and rename work_processor_pipelines
session property to late_materialization
. (#2275)
Hive Connector Changes
- Fix incorrect query results when reading
timestamp
values from ORC files written by
Hive 3.1 or later. (#2099)
- Fix a CDH 5.x metastore compatibility issue resulting in failure when analyzing or inserting
into a table with
date
columns. (#556)
- Reduce number of metastore calls when fetching partitions. (#1921)
- Support reading from insert-only transactional tables. (#576)
- Deprecate
parquet.fail-on-corrupted-statistics
(previously known as hive.parquet.fail-on-corrupted-statistics
).
Setting this configuration property to false
may hide correctness issues, leading to incorrect query results.
Session property parquet_fail_with_corrupted_statistics
is deprecated as well.
Both configuration and session properties will be removed in a future version. (#2129)
- Improve concurrency when updating table or partition statistics. (#2154)
- Add support for renaming views. (#2189)
- Allow configuring the
hive.orc.use-column-names
config property on a per-session
basis using the orc_use_column_names
session property. (#2248)
Kudu Connector Changes
- Support predicate pushdown for the
decimal
type. (#2131)
- Fix column position swap for delete operations that may result in deletion of the wrong records. (#2252)
- Improve predicate pushdown for queries that match a column against
multiple values (typically using the
IN
operator). (#2253)
MongoDB Connector Changes
- Add support for reading from views. (#2156)
PostgreSQL Connector Changes
- Allow converting unsupported types to
VARCHAR
by setting the session property
unsupported_type_handling
or configuration property unsupported-type-handling
to CONVERT_TO_VARCHAR
. (#1182)
MySQL Connector Changes
- Fix
INSERT
query failure when GTID
mode is enabled. (#2251)
Elasticsearch Connector Changes
- Improve performance for queries involving equality and range filters
over table columns. (#2310)
Google Sheets Connector Changes
- Fix incorrect results when listing tables in
information_schema
. (#2118)
SPI Changes
- Add
executionTime
to QueryStatistics
for event listeners. (#2247)