Release 332 (08 Apr 2020)#


  • Fix query failure during planning phase for certain queries involving multiple joins. (#3149)

  • Fix execution failure for queries involving large IN predicates on decimal values with precision larger than 18. (#3191)

  • Fix prepared statements or view creation for queries containing certain nested aliases or TABLESAMPLE clauses. (#3250)

  • Fix rare query failure. (#2981)

  • Ignore trailing whitespace when loading configuration files such as etc/ or etc/ Trailing whitespace in etc/ and catalog properties files was already ignored. (#3231)

  • Reduce overhead for internal communication requests. (#3215)

  • Include filters over all table columns in output of EXPLAIN (TYPE IO). (#2743)

  • Support configuring multiple event listeners. The properties files for all the event listeners can be specified using the event-listener.config-files configuration property. (#3128)

  • Add CREATE SCHEMA ... AUTHORIZATION syntax to create a schema with specified owner. (#3066).

  • Add optimizer.push-partial-aggregation-through-join configuration property to control pushing partial aggregations through inner joins. Previously, this was only available via the push_partial_aggregation_through_join session property. (#3205)

  • Rename configuration property optimizer.push-aggregation-through-join to optimizer.push-aggregation-through-outer-join. (#3205)

  • Add operator statistics for the number of splits processed with a dynamic filter applied. (#3217)


  • Fix LDAP authentication when user belongs to multiple groups. (#3206)

  • Verify access to table columns when running SHOW STATS. (#2665)

  • Only return views accessible to the user from information_schema.views. (#3290)

JDBC driver#

  • Add clientInfo property to set extra information about the client. (#3188)

  • Add traceToken property to set a trace token for correlating requests across systems. (#3188)

BigQuery connector#

  • Extract parent project ID from service account before looking at the environment. (#3131)

Elasticsearch connector#

  • Add support for ip type. (#3347)

  • Add support for keyword fields with numeric values. (#3381)

  • Remove unnecessary configuration property. (#3265)

Hive connector#

  • Fix failure reading certain Parquet files larger than 2GB. (#2730)

  • Improve performance when reading gzip-compressed Parquet data. (#3175)

  • Explicitly disallow reading from Delta Lake tables. Previously, reading from partitioned tables would return zero rows, and reading from unpartitioned tables would fail with a cryptic error. (#3366)

  • Add configuration property for setting the permissions of new directories created by Presto. Default value is 0777, which corresponds to previous behavior. (#3126)

  • Add hive.partition-use-column-names configuration property and matching partition_use_column_names catalog session property that allows to match columns between table and partition schemas by names. By default they are mapped by index. (#2933)

  • Add support for CREATE SCHEMA ... AUTHORIZATION to create a schema with specified owner. (#3066).

  • Allow specifying the Glue metastore endpoint URL using the hive.metastore.glue.endpoint-url configuration property. (#3239)

  • Add experimental file system caching. This can be enabled with the hive.cache.enabled configuration property. (#2679)

  • Support reading files compressed with newer versions of LZO. (#3209)

  • Add support for Alluxio Catalog Service. (#2116)

  • Remove unnecessary hive.metastore.glue.use-instance-credentials configuration property. (#3265)

  • Remove unnecessary hive.s3.use-instance-credentials configuration property. (#3265)

  • Add flexible S3 security mapping, allowing for separate credentials or IAM roles for specific users or buckets/paths. (#3265)

  • Add support for specifying an External ID for an IAM role trust policy using the hive.metastore.glue.external-id configuration property (#3144)

  • Allow using configured S3 credentials with IAM role. Previously, the configured IAM role was silently ignored. (#3351)

Kudu connector#

  • Fix incorrect column mapping in Kudu connector. (#3170, #2963)

  • Fix incorrect query result for certain queries involving IS NULL predicates with OR. (#3274)

Memory connector#

  • Include views in the list of tables returned to the JDBC driver. (#3208)

MongoDB connector#

  • Add objectid_timestamp for extracting the timestamp from ObjectId. (#3089)

  • Delete document from _schema collection when DROP TABLE is executed for a table that exists only in _schema. (#3234)

SQL Server connector#

  • Disallow renaming tables between schemas. Previously, such renames were allowed but the schema name was ignored when performing the rename. (#3284)


  • Expose row filters and column masks in QueryCompletedEvent. (#3183)

  • Expose referenced functions and procedures in QueryCompletedEvent. (#3246)

  • Allow Connector to provide EventListener instances. (#3166)

  • Deprecate the ConnectorPageSourceProvider.createPageSource() variant without the dynamicFilter parameter. The method will be removed in a future release. (#3255)