先从app表提取出用户的uid,触达页面的时间time,触达页面tab
然后使用使用collect_set()将所有页面组合起来,并通过逗号连接
collect_list()同理
select uid,concat_ws(’,’, collect_set(tab))
from
(select uid,time,tab from app
group by uid,time,tab
order by uid,time ) aa
group by uid
collect_set()和collect_list()的区别
collect_set()提取的结果是去重的结果,即用户访问了哪些不重复的页面
collect_set()提取的结果是以时间为顺序用户访问了哪些页面,包括重复访问