hive> select ti,count(1) from xx where day=20150330 group by ti > union all
> select ti,count(1) from xx where day=20150330 group by ti;Total jobs =
3Launching Job 1 out of 3
From: smart...@hotmail.com
To: user@hive.apache.org
Subject: hive union all same table
Date: Wed, 1 Apr 2015 03:06:52 +0000
Hi I want to query table_a by different where clause as follow
select * from ( select col1, count(1) from table_a where col2 ='xx'
group by col1 union all select col1, count(1) from table_a
where col2 ='xx' group by col1 union all select col1,
count(1) from table_a where col2 ='xx' group by col1 )
I found the total jobs is 4， is there any optimiztion? I think it should scan
table_a once and only one job
