hive常用的一些函数

1.分组后查询前几列数据

row_number() over (partition by p_day,uid order by time asc) num

num = 1 取第一个,num<=10取分区后前十个

2.lag/lead 某一行前/后附近一行的数据

lag(url,1,2001) over (partition by uid order by time) url_1

url这一列2001这一行前一列数据

猜你喜欢

转载自www.cnblogs.com/haoziii/p/10132720.html