Apache Hive 1.2 drops with new SQL functions
A total of approximately 480 JIRA tickets is what it takes to update Apache Hive to Version 1.2. The data warehouse software for Apache Hadoop has already reached its third release of the year, with the Hive community continuing its growth.
The third major update for Apache Hive this year brings primarily new SQL features as well as improvements in the areas of security, performance, and stability. The developers behind the new shipment have highlighted the aim of the new changes as being concerned with making Apache Hive ready for enterprise deployment.
The new SQL functions include support of the union command (previously only union all was supported), with added support for specifying column list in insert statement. To increase the safety, improvements were made to the authorisation plugin API, whereby implementations for filtering the results of metadata operations are possible.
SEE ALSO: What’s new in Apache Hive 1.0.0?
Finally, HiveServer2 HTTP transport mode is now supported for cookie-based authentication. Advances in performance and optimisation consist of statistics caching in HiveServer2 and an improved hash join algorithm. In terms of usability, an improvement to HS2 logging is there, allowing logging verbosity to be set at session level.
Sushanth Sowmyan and Wei Wang were proud to announce the Hive developer community growth in tandem with the newest Hive release:
SQL is the most popular use case for the Hadoop user community, and Apache Hive is still the defacto standard. Early this week, the Apache Hive community released Apache Hive 1.2.0. Already the third release this year, the Hive developer community continues to improve the release and grow its team, with 11 Hive contributors promoted to committers in the last three months.
The Apache Hive release schedule recently became more fast-paced, with the community working at having a more incremental release timetable. Although Version 1.2 is being labelled as having major improvements for Enterprise SQL, the release is still smaller than Hive 1.1, which featured Hive-on-Spark.
A somewhat incomplete list of changes can be found in the detailed release notes. Most of the Hive-on-Spark JIRAs are missing from this list.