当前位置:首页 > 数据库 > 正文内容

sql查询用户连续n天登录的数据

2024-11-29数据库3

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。

1
2
3
4
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。

1
2
3
4
5
6
7
8
9
10
11
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。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
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:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
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