[IMPALA-1270] Consider adding distinct aggregation to subqueries as perf optimization

• Type: Task
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: Impala 2.0
• Fix Version/s: Impala 4.0.0
• Component/s: Frontend
• Labels: planner
• Epic Link: Subquery Support
• Target Version: Product Backlog
• Assignee: Tim Armstrong
• Reporter: Nong Li
• Created: 18/Sep/14 20:12
• Updated: 15/Dec/20 19:57
• Resolved: 15/Jul/20 18:00

Description

We should consider other rewrites for exists. For q4, another rewrite is an inner join + distinct:

select
  o_orderpriority,
  count(distinct l_orderkey) as order_count
from lineitem l
inner join orders o
  on (o.o_orderkey = l.l_orderkey and
      l.l_commitdate < l.l_receiptdate)
where
  o_orderdate >= '1993-07-01' and
  o_orderdate < '1993-10-01'
group by
  o_orderpriority
order by
  o_orderpriority

This can run much faster because we have more flexibility on how we execute the inner join. We get killed partitioning lineitem now.

Issue Links

• is duplicated by: IMPALA-1728 sub-query with duplicate values used IN conditional operator should discard the duplicate values before applying the operator