Release 358 (1 Jun 2021)#
General#
Support arbitrary queries in SHOW STATS. (#8026)
Improve performance of complex queries involving joins and
TABLESAMPLE
. (#8094)Improve performance of
ORDER BY ... LIMIT
queries on sorted data. (#6634)Reduce graceful shutdown time for worker nodes. (#8149)
Fix query failure columns of non-orderable types (e.g.
HyperLogLog
,tdigest
, etc.), are involved in a join. (#7723)Fix failure for queries containing repeated ordinals in a
GROUP BY
clause. Example:SELECT x FROM t GROUP BY 1, 1
. (#8023)Fix failure for queries containing repeated expressions in the
ORDER BY
clause of an aggregate function. Example:SELECT array_agg(x ORDER BY y, y) FROM (VALUES ('a', 2)) t(x, y)
. (#8080)
JDBC Driver#
Hive connector#
Iceberg connector#
Fix failure when listing materialized views in
information_schema.tables
or via thejava.sql.DatabaseMetaData.getTables()
JDBC API. (#8151)
Memory connector#
Improve performance of certain complex queries involving joins. (#8095)
SPI#
Remove deprecated
ConnectorPageSourceProvider.createPageSource()
method overrides. (#8077)Add support for casting the columns of a redirected table scan when source column types don’t match. (#6066)
Add
ConnectorMetadata.redirectTable()
to allow connectors to redirect table reads and metadata listings. (#7606)Add
ConnectorMetadata.streamTableColumns()
for streaming column metadata in a redirection-aware manner. The alternate method for column listingConnectorMetadata.listTableColumns()
is now deprecated. (#7606)