Release 301 (31 Jan 2019)#

General#

  • Fix reporting of aggregate input data size stats. (#100)

  • Add support for role management (see CREATE ROLE). Note, using SET ROLE requires an up-to-date client library. (#90)

  • Add INVOKER security mode for CREATE VIEW. (#30)

  • Add ANALYZE SQL statement for collecting table statistics. (#99)

  • Add log() function with arbitrary base. (#36)

  • Remove the deprecated.legacy-log-function configuration option. The legacy behavior (reverse argument order) for the log() function is no longer available. (#36)

  • Remove the deprecated.legacy-array-agg configuration option. The legacy behavior (ignoring nulls) for array_agg() is no longer available. (#77)

  • Improve performance of COALESCE expressions. (#35)

  • Improve error message for unsupported reduce_agg() state type. (#55)

  • Improve performance of queries involving SYSTEM table sampling and computations over the columns of the sampled table. (#29)

Server RPM#

  • Do not allow uninstalling RPM while server is still running. (#67)

Security#

  • Support LDAP with anonymous bind disabled. (#97)

Hive connector#

  • Add procedure for dumping metastore recording to a file. (#54)

  • Add Metastore recorder support for Glue. (#61)

  • Add hive.temporary-staging-directory-enabled configuration property and temporary_staging_directory_enabled session property to control whether a temporary staging directory should be used for write operations. (#70)

  • Add hive.temporary-staging-directory-path configuration property and temporary_staging_directory_path session property to control the location of temporary staging directory that is used for write operations. The ${USER} placeholder can be used to use a different location for each user (e.g., /tmp/${USER}). (#70)

Kafka connector#

  • The minimum supported Kafka broker version is now 0.10.0. (#53)

Base-JDBC connector library#

  • Add support for defining procedures. (#73)

  • Add support for providing table statistics. (#72)

SPI#

  • Include session trace token in QueryCreatedEvent and QueryCompletedEvent. (#24)

  • Fix regression in NodeManager where node list was not being refreshed on workers. (#27)