hive去重操作

1、建表

create table lintest(id string,name string,size int);

2、插入数据

create table inform(id string,name string,size int);

3、去重查询

select ad ,sum(plus),count(distinct name,id)
from 
(select distinct name,id,ad,plus from lintest) tmp
group by ad

4、求和之前不去重

select ad,sum(plus),count(distinct name,id) from lintest group by ad;

5、row_number()去重

select uuid,build_brand,substring(os,0,9) as version,num from(
select uuid,build_brand,os,row_number() over(distribute by uuid sort by build_brand) as num
from base.user_all)tmp
where tmp.num=1