row_number()函数

row_number() over(partition by col1 order by col2 desc) rank
按照col1分组,col2字段组内排序

语法:row_number() over (partition by 字段a order by 计算项b desc ) rank 
--这里rank是别名

partition by:类似hive的建表,分区的意思;
order by :排序,默认是升序,加desc降序;
这里按字段a分区,对计算项b进行降序排序

实例:
要取top10品牌,各品牌的top10渠道,各品牌的top10渠道中各渠道的top10档期 
1、取top10品牌
select 品牌,count/sum/其它() as num  from  table_name  order by num limit 10;

 2、 取top10品牌下各品牌的top10渠道        
select 
          a.*
from
          (
                  select 品牌,渠道,count/sum/其它() as num row_number() over (partition by 品牌 order by num desc ) rank  
                   from table_name
                   where 品牌限制条件
                   group by 品牌,渠道
          )a
where 
          a.rank<=10
 3、 取top10品牌下各品牌的top10渠道中各渠道的top10档期
 
select 
          a.*
from
          (
                  select 品牌,渠道,档期,count/sum/其它() as num row_number() over (partition by 品牌,渠道 order by num desc ) rank  
                   from table_name
                   where 品牌,渠道 限制条件
                   group by 品牌,渠道,档期
          )a
where 
          a.rank<=10

猜你喜欢

转载自blog.csdn.net/java_soldier/article/details/80108551