hive去重方式
1.distinct 去重字段列表
特点：对distinct后面的字段列表去重，无参考系
select distinct case_id, role, judgename
from judgeInfo;
2.row_number() over(partition by 去重字段列表 order by 参考系字段)
特点：有参考系，即对某字段排序，取序号为特点值的记录。
如，去重，取最新记录：
select case_id, role, judgename,dt from
(select case_id, role, judgename,dt,
row_number() over(partition by case_id, role, judgename order by dt desc) rnk
from judgeInfo) a
where rnk=1;