问题
nnnn有如下访问网站的数据,包括用户id和访问时间两个字段。如果某个用户的连续的访问记录时间间隔小于60秒,则属于同一个会话,现在需要计算每个用户有多少个会话。
nnnn比如A用户在第1秒,60秒,200秒,230秒有三次访问记录,则该用户有2个会话,其中第一个会话是第1秒和第60秒的记录,第二个会话是第200秒和230秒的记录。
nnnnuser_id tsn1001 16920000000n1001 16920000050n1002 16920000065n1002 16920000080n1001 16920000150n1002 16920000160
nnnn解答
nnnnuser_id | ts | 判断与上一行差值是否小于60 | 开窗累加当做会话编号 |
---|---|---|---|
A | 1 | 0 | 0 |
A | 60 | 0 | 0 |
A | 200 | 1 | 1 |
A | 230 | 0 | 1 |
with tmp as (n select 1001 as user_id,16920000000 as tsn union alln select 1001 as user_id,16920000050 as tsn union alln select 1002 as user_id,16920000065 as tsn union alln select 1002 as user_id,16920000080 as tsn union alln select 1001 as user_id,16920000150 as tsn union alln select 1002 as user_id,16920000160 as tsn)n nselectn user_idn ,count(distinct user_group) as user_group_cntnfromn(n selectn user_idn ,tsn -- 开窗做累加n ,sum(flag) over(partition by user_id order by ts) as user_groupn fromn (nselectn user_idn ,tsn -- 判断当前行的时间与上一行的差值n ,if(ts-last_ts<60,0,1) as flagn fromn (nselectn user_idn ,tsn -- 取当前行的上一个时间,没有上一行就给自身的时间n ,lag(ts,1,ts) over(partition by user_id order by ts) as last_tsn from tmpn)t1n)t1n)t1ngroup by user_id;
n
Comments NOTHING