Apache Hive 0.12: Stinger Phase Two… DELIVERED by Thejas Nair.
From the post:
Stinger is not a product. Stinger is a broad community based initiative to bring interactive query at petabyte scale to Hadoop. And today, as representatives of this open, community led effort we are very proud to announce delivery of Apache Hive 0.12, which represents the critical second phase of this project!
Only five months in the making, Apache Hive 0.12 comprises over 420 closed JIRA tickets contributed by ten companies, with nearly 150 thousand lines of code! This work is perfectly representative of our approach… it is a substantial release with major contributions from a wide group of talented engineers from Microsoft, Facebook , Yahoo and others.
Delivery of SQL-IN-Hadoop Marches
The Stinger Initiative was announced in February and as promised, we have seen consistent regular delivery of new features and improvements as outlined in the Stinger plan. There are three roadmap vectors for Stinger: Speed, Scale and SQL. Each phase of the initiative advances on all three goals and this release provides a significant increase in SQL semantics, adding the
VARCHARandDATEdatatypes and improving performanceORDERby andGROUPby. Several features to optimize queries have also been added.We also contributed numerous “under the hood” improvements, ie refactoring code and making it easier to build on top of hive – getting rid of some of the technical debt. This helps us deliver further optimizations in the long term, especially for the upcoming Apache Tez integration.
A complete list of the notable improvements included in the release is listed here and expect an updated performance benchmark soon!
It is so nice to see a successful software project!
And an open source one at that!
Unlike the no bid IT mega-failure that is Obamacare.
Maybe there is something to having a good infrastructure for code development as opposed to contractors billing by the phone call, lunch meeting and hour.
BTW, all the protests about the volume of users trying to register with Obamacare? More managerial incompetence.
When you are rolling out a system for potentially 300 million+ users, don’t you anticipate load as part of the requirements?
If you didn’t, there is the start of the trail of managerial incompetence in Obamacare.