Stamatis Zampetakis created HIVE-28829:
------------------------------------------
Summary: Remove redundant aggregate operator from query plan
Key: HIVE-28829
URL: https://issues.apache.org/jira/browse/HIVE-28829
Project: Hive
Issue Type: Improvement
Components: CBO
Reporter: Stamatis Zampetakis
Assignee: Stamatis Zampetakis
Some queries contain redundant aggregate operators in the query plan. One such
example is shown below.
{code:sql}
create table t1 (a1 int, b1 string);
explain cbo
select b1 from t1 where a1 = 1
union
select b1 from t1 where a1 = 2
union
select b1 from t1 where 1 = 0;
{code}
{noformat}
HiveAggregate(group=[{0}])
HiveProject(b1=[$0])
HiveAggregate(group=[{0}])
HiveProject(b1=[$0])
HiveUnion(all=[true])
HiveProject(b1=[$1])
HiveFilter(condition=[=($0, 1)])
HiveTableScan(table=[[default, t1]], table:alias=[t1])
HiveProject(b1=[$1])
HiveFilter(condition=[=($0, 2)])
HiveTableScan(table=[[default, t1]], table:alias=[t1])
{noformat}
It's pretty obvious that the top {{HiveAggregate}} and {{HiveProject}}
operators are completely redundant and can be removed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
