Pig on Tez work is very active and proceeding well, and I would expect it to part of a release in couple of months (less than 6 months).
There is also now better support for ORC format in pig as well. The major stinger aspect that pig hasn’t benefited from is the vectorization changes. Other improvements like Tez and ORC will be part of pig.
Since these technologies are constantly improving, I would reccomend choosing what fits well for your current use cases. This means also looking at if you want the expressibility of pig and also the conciseness of SQL.
Hive / HCatalog Forum
Hive vs Pig – Performance
I have a scenario similar to ETL and i was inclined towards using Pig. Most of the comparisons which i have seen/read about showcase a performance gain of 20-40% using Pig. However with the completion of Stinger project in HDP 2.1, I would like to know how Pig and Hive stack up in terms of performance. Pig on Tez is a future roadmap and is not yet implemented. Atleast till this time I beleieve Hive should be faster that Pig. Please let me know if any of you have any information on this.
Support from the Experts
A HDP Support Subscription connects you experts with deep experience running Apache Hadoop in production, at-scale on the most demanding workloads.
Become HDP Certified
Real world training designed by the core architects of Hadoop. Scenario-based training courses are available in-classroom or online from anywhere in the world