Release 331 (16 Mar 2020)#

General#

  • Prevent query failures when worker is shut down gracefully. (#2648)

  • Fix join failures for queries involving OR predicate with non-comparable functions. (#2861)

  • Ensure query completed event is fired when there is an error during analysis or planning. (#2842)

  • Fix memory accounting for ORDER BY queries. (#2612)

  • Fix last_day_of_month() for timestamp with time zone values. (#2851)

  • Fix excessive runtime when parsing deeply nested expressions with unmatched parenthesis. (#2968)

  • Correctly reject date literals that cannot be represented in Presto. (#2888)

  • Improve query performance by removing redundant data reshuffling. (#2853)

  • Improve performance of inequality joins involving BETWEEN. (#2859)

  • Improve join performance for dictionary encoded data. (#2862)

  • Enable dynamic filtering by default. (#2793)

  • Show reorder join cost in EXPLAIN ANALYZE VERBOSE (#2725)

  • Allow configuring resource groups selection based on user’s groups. (#3023)

  • Add SET AUTHORIZATION action to ALTER SCHEMA. (#2673)

  • Add BigQuery коннектор. (#2532)

  • Add support for large prepared statements. (#2719)

Security#

JDBC driver#

  • Implement PreparedStatement.getMetaData(). (#2770)

Web UI#

  • Fix copying worker address to clipboard. (#2865)

  • Fix copying query ID to clipboard. (#2872)

  • Fix display of data size values. (#2810)

  • Fix redirect from / to /ui/ when Presto is behind a proxy. (#2908)

  • Fix display of prepared queries. (#2784)

  • Display physical input read rate. (#2873)

  • Add simple form based authentication that utilizes the configured password authenticator. (#2755)

  • Allow disabling the UI via the web-ui.enabled configuration property. (#2755)

CLI#

  • Fix formatting of varbinary in nested data types. (#2858)

  • Add --timezone parameter. (#2961)

Hive connector#

  • Fix incorrect results for reads from information_schema tables and metadata queries when using a Hive 3.x metastore. (#3008)

  • Fix query failure when using Glue metastore and the table storage descriptor has no properties. (#2905)

  • Fix deadlock when Hive caching is enabled and has a refresh interval configured. (#2984)

  • Respect bucketing_version table property when using Glue metastore. (#2905)

  • Improve performance of partition fetching from Glue. (#3024)

  • Add support for bucket sort order in Glue when creating or updating a table or partition. (#1870)

  • Add support for Hive full ACID tables. (#2068, #1591, #2790)

  • Allow data conversion when reading decimal data from Parquet files and precision or scale in the file schema is different from the precision or scale in partition schema. (#2823)

  • Add option to enforce that a filter on a partition key be present in the query. This can be enabled by setting the hive.query-partition-filter-required configuration property or the query_partition_filter_required session property to true. (#2334)

  • Allow selecting the Intelligent-Tiering S3 storage class when writing data to S3. This can be enabled by setting the hive.s3.storage-class configuration property to INTELLIGENT_TIERING. (#3032)

  • Hide the Hive system schema sys for security reasons. (#3008)

  • Add support for changing the owner of a schema. (#2673)

MongoDB connector#

  • Fix incorrect results when queries contain filters on certain data types, such as real or decimal. (#1781)

Other connectors#

These changes apply to the MemSQL, MySQL, PostgreSQL, Redshift, Phoenix, and SQL Server connectors.

  • Add support for dropping schemas. (#2956)

SPI#

  • Remove deprecated Identity constructors. (#2877)

  • Introduce a builder for ConnectorIdentity and deprecate its public constructors. (#2877)

  • Add support for row filtering and column masking via the getRowFilter() and getColumnMask() APIs in SystemAccessControl and ConnectorAccessControl. (#1480)

  • Add access control check for executing procedures. (#2924)