业务中常见分析之一是分析用户连续登录使用情况,这也对应着SQL常见面试题——用户连续N天登录问题。
我们假设现在有一张用户登录信息表user_login_info,表中字段有用户id(uid)、登录时间(login_time)。表中数据如下所示:
现在要求查询出连续登录N天的用户。
(1) 首先我们要对用户登录表进行去重操作,以避免用户当天多次登录情况对查询结果产生影响。这里用到了distinct关键词。
select
uid,
distinct date(login_time) as login_time
from user_login_info
(2)其次我们使用窗口排名函数row_number对同一用户的不同登录时间进行排名,得到新一列为rk。
select
uid,
login_time,
row_number() over(partition by uid order by login_time) as rk
from
(
select
uid,
distinct date(login_time) as login_time
from user_login_info
) t1
查询结果如图所示:
(3)之后用date_sub函数计算登录时间login_time一列加上rk天之后生成新的一列sub_date。假如在表格中,同一用户的sub_date相同则说明相同sub_date数据的行是连续登录使用的情况。
select *, DATE_SUB(login_time, INTERVAL rk DAY) as sub_date
from
(
select
uid,
login_time,
row_number() over(partition by uid order by login_time) as rk
from
(
select
uid,
distinct date(login_time) as login_time
from user_login_info
) t1
) t2
查询结果如图所示:
(4)之后我们对得到的查询结果,按照用户id,登录时间进行分组计数,得到的计数结果就是用户连续登录多少天的情况记录。
select uid, count(*) as 连续登录天数
from
(
SELECT *, DATE_SUB(login_time, INTERVAL rk DAY) AS sub_date
from
(
select
uid,
login_time,
row_number() over(partition by uid order by login_time) as rk
from
(
select
uid,
distinct date(login_time) as login_time
from user_login_info
) t1
) t2
) t3
group by uid, sub_date
(5)之后我们可以在此查询结果上,根据需要用having条件就可以筛选出我们想要得知的连续N天登录的用户id
完整代码如下:
select uid, count(*) as 连续登录天数
from
(
SELECT *, DATE_SUB(login_time, INTERVAL rk DAY) AS sub_date
from
(
select
uid,
login_time,
row_number() over(partition by uid order by login_time) as rk
from
(
select
uid,
distinct date(login_time) as login_time
from user_login_info
) t1
) t2
) t3
group by uid, sub_date
having 连续登录天数 = N
到此这篇关于SQL查询用户连续N天登录的文章就介绍到这了,更多相关SQL查询连续N天登录内容请搜索编程网(www.lsjlt.com)以前的文章或继续浏览下面的相关文章希望大家以后多多支持编程客栈(www.lsjlt.com)!