11.122. Release 0.119
- Add Redis Connector.
- Add function.
- Fix restoring interrupt status in
StatementClient
. - Support getting server version in JDBC driver.
- Catalog and schema are now optional on the server. This allows connectingand executing metadata commands or queries that use fully qualified names.Previously, the CLI and JDBC driver would use a catalog and schema named
default
if they were not specified. - Fix scheduler handling of partially canceled queries.
- Execute views with the permissions of the view owner.
- Replaced the
task.http-notification-threads
config option with twoindependent options: andtask.http-timeout-threads
. - Improve handling of negated expressions in join criteria.
- Fix
arbitrary()
, andmin_by()
functions when usedwith an array, map or row type. - Fix union coercion when the same constant or column appears more than once onthe same side.
- Support in .
SPI Changes
- Add more system table distribution modes.
- Add owner to view metadata.
These are backwards incompatible changes with the previous connector SPI.If you have written a connector, you may need to update your code to thenew APIs.
- Fix handling of full width characters.
- Allow performing a partial query cancel using
ctrl-P
. - Allow toggling debug mode during query by pressing
D
. - Fix handling of query abortion after result has been partially received.
- Fix handling of
ctrl-C
when displaying results without a pager.
Verifier Changes
- Add
expected-double-precision
config to specify the expected level ofprecision when comparing double values. - Return non-zero exit code when there are failures.
- Add support for Cassandra blob types.
Hive Changes
- Support adding and renaming columns using .
- Automatically configure the S3 region when running in EC2.
- Allow configuring multiple Hive metastores for high availability.
- Add support for and
VARBINARY
in Parquet.
- Fix handling of pattern characters when matching table or column names.