Important new features of Apache Hive in Hortonworks Data Platform (HDP) 2.6 include the following:

Full Support and Easier Setup for Interactive SQL Queries

Hortonworks supports interactive SQL queries with Hive low-latency analytical processing (LLAP). In the previous HDP release, Hive LLAP was a Technical Preview. Persistent query infrastructure and optimized data caching enable Hive LLAP. See Setting Up Hive LLAP for how to enable Hive LLAP in Apache Ambari.

LLAP-Focused Revision of the HDP Apache Hive Performance Tuning Guide

The Hive Performance Tuning Guide focuses on Hive LLAP to leverage gains in performant data analytics technologies. The guide includes a diverse set of recommendations that are tailored to optimizing the Tez framework for efficient querying of a Hive EDW. While interactive querying is emphasized, the documentation includes guidelines for environments that prefer or must use Hive on Tez without LLAP.

Support for the ACID-Transaction MERGE Statement

Hive in HDP 2.6 supports the MERGE SQL statement, which simplifies updating data, deleting data, and change data captures among tables.

Enhancements for SQL and TPC-DS Completeness

You can offload SQL workload to Hive using all TPC-DS queries with only trivial rewrites. The addition of multiple and scalar subqueries, INTERSECT and EXCEPT operators, standard syntax for ROLLUP / GROUPING SET, and syntax improvements for GROUP BY and ORDER BY make the SQL offloading easier than before.

Hive View 2.0

With the release of Apache Ambari 2.5.0, HDP provides Hive View 2.0. Hive View 2.0 is an enhancement over earlier Hive View UIs because it has more nimble ways to create and manage databases and tables, offers visual explain plans of your queries that pinpoint workload costs, and provides a way to view security policies at the table level.