oracle查询连续n天登录的用户

最新推荐文章于 2024-07-25 12:01:43 发布

weixin_30758821

最新推荐文章于 2024-07-25 12:01:43 发布

阅读量483

点赞数 1

原文链接：http://www.cnblogs.com/crazyleeyang/p/3892399.html

版权

-- 查询连续3天登录的用户

1 先创建一个表,如下:

1 create table USER_DATA
2 (
3 USER_ID NUMBER,
4 LOGIN_TIME DATE
5 );

2 插入用户登录数据:

 1 insert into user_data (USER_ID, LOGIN_TIME) values ('10002', to_date('21-07-2014', 'dd-mm-yyyy'));
 2 insert into user_data (USER_ID, LOGIN_TIME) values ('10002', to_date('22-07-2014', 'dd-mm-yyyy'));
 3 insert into user_data (USER_ID, LOGIN_TIME) values ('10002', to_date('24-07-2014', 'dd-mm-yyyy'));
 4 insert into user_data (USER_ID, LOGIN_TIME) values ('10002', to_date('25-07-2014', 'dd-mm-yyyy'));
 5 insert into user_data (USER_ID, LOGIN_TIME) values ('10003', to_date('21-07-2014', 'dd-mm-yyyy'));
 6 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('28-07-2014', 'dd-mm-yyyy'));
 7 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('29-07-2014', 'dd-mm-yyyy'));
 8 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('30-07-2014', 'dd-mm-yyyy'));
 9 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('21-07-2014', 'dd-mm-yyyy'));
10 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('22-07-2014', 'dd-mm-yyyy'));
11 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('23-07-2014', 'dd-mm-yyyy'));
12 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('24-07-2014', 'dd-mm-yyyy'));
13 insert into user_data (USER_ID, LOGIN_TIME) values ('10001', to_date('26-07-2014', 'dd-mm-yyyy'));

3 查询连续3天登录的用户

思路:现有的数据里面已经是经过处理的,每个用户每天只有一条数据,并且登录时间是每天的零点,如果是真实的数据则进行trunc后就行,然后选择的时候,需要依据user_id和login_time进行排序,这样在外层选择时,rownm在每个用户按日志排序时是连续的,然后查询时用日志减去rownum,这样如果是连续的日期,其相减的结果就是一样的,然后根据此结果来判断是否是连续登录就行了,查询连续3天登录的用户id的具体代码如下:

 1 select distinct user_id
 2 from
 3 (
 4     select b.user_id, b.d_temp, count(*)
 5     from
 6     (
 7         select a.user_id, a.login_time - rownum d_temp
 8         from
 9         (
10             select t.*
11             from user_data t
12             order by t.user_id, t.login_time
13         )a
14     )b
15     group by b.user_id, b.d_temp
16     having count(*) >= 3
17 );