Public signup for this instance is disabled. Our Jira Guidelines page explains how to get an account.
select count(distinct ss_ticket_number) from store_sales;
can be rewritten as
select count(1) from (select distinct ss_ticket_number from store_sales) a;
which may run upto 3x faster
Combination of ReducesinkDedup + TopN optimization yields incorrect result if there are multiple GBY in reducer
Make HIVE-10568 work with Spark [Spark Branch]
Select count(distinct()) a couple of times stuck in last reducer