Release 0.186#

Предупреждение

This release has a stability issue that may cause query failures in large deployments due to HTTP requests timing out.

General#

  • Fix excessive GC overhead caused by map to map cast.

  • Fix implicit coercions for ROW types, allowing operations between compatible types such as ROW(INTEGER) and ROW(BIGINT).

  • Fix issue that may cause queries containing expensive functions, such as regular expressions, to continue using CPU resources even after they are killed.

  • Fix performance issue caused by redundant casts.

  • Fix json_parse() to not ignore trailing characters. Previously, input such as [1,2]abc would successfully parse as [1,2].

  • Fix leak in running query counter for failed queries. The counter would increment but never decrement for queries that failed before starting.

  • Reduce coordinator HTTP thread usage for queries that are queued or waiting for data.

  • Reduce memory usage when building data of VARCHAR or VARBINARY types.

  • Estimate memory usage for GROUP BY more precisely to avoid out of memory errors.

  • Add queued time and elapsed time to the client protocol.

  • Add query_max_execution_time session property and query.max-execution-time config property. Queries will be aborted after they execute for more than the specified duration.

  • Add inverse_normal_cdf() function.

  • Add Geospatial functions including functions for processing Bing tiles.

  • Add Spill (выгрузка данных на диск) for joins.

  • Add Redshift коннектор.

Resource groups#

  • Query Queues are deprecated in favor of Группы ресурсов and will be removed in a future release.

  • Rename the maxRunning property to hardConcurrencyLimit. The old property name is deprecated and will be removed in a future release.

  • Fail on unknown property names when loading the JSON config file.

JDBC driver#

  • Allow specifying an empty password.

  • Add getQueuedTimeMillis() and getElapsedTimeMillis() to QueryStats.

Hive#

  • Fix FileSystem closed errors when using Kerberos authentication.

  • Add support for path style access to the S3 file system. This can be enabled by setting the hive.s3.path-style-access=true config property.

SPI#

  • Add an ignoreExisting flag to ConnectorMetadata::createTable().

  • Remove the getTotalBytes() method from RecordCursor and ConnectorPageSource.

Примечание

These are backwards incompatible changes with the previous SPI. If you have written a connector, you will need to update your code before deploying this release.