Community Articles

Find and share helpful community-sourced technical articles.
Labels (2)
avatar
Rising Star

So I might be late in posting this article, but I want to publicize the changes in HDP 2.6.0 documentation and blog postings for Hive, Hive Views, and Ambari that were released in early April 2017. After all, I don't think all of you out there are upgraded to HDP 2.6.0 yet.

To complement the G.A. release of Hive low-latency analytical processing (LLAP), the "Hive Performance Tuning Guide" has been completely rewritten to focus on how to optimize your environment and queries for interactive processing. While the guide is tailored to run a Hive data warehouse on the Tez framework, the new edition still has relevant tips and links to information for users who do batch processing on a MapReduce framework.

A couple of LLAP blogs to consider reading are "Fast Analytics in the Cloud with Hive LLAP" and "Top 5 Performance Boosters with Apache Hive LLAP".

Important debugging and optimization tools to get the most out of your Hive queries and your architecture are documented in Hive View 2.0 and Tez View, both of which introduce improvements over earlier versions of Ambari Views. You might also want to check out the "3 Great Reasons to Try Apache Hive View 2.0" blog.

SQL-standard ACID MERGE support in HDP 2.6 raises the bar for Hive transactional data management by adding a helpful capability to ACID tables in HDP Hive, which already had supported record-level INSERT, UPDATE, and DELETE operations within tables. The current plan is to expand the documentation to include a step-by-step procedure for creating an ACID table with Apache Ambari. In the meantime, a good overview of why to use this functionality is in the "Apache Hive: Moving Beyond Analytics Offload with SQL MERGE" blog.

2,219 Views
Comments

@Frank Welsch the LLAP blogs links back to your HCC article.

@Scott Shaw

Thanks for alerting me! I've fixed this, and I think everything should be in order now. Please let me know if you find anything else askew.