Question

我无法避免此查询中的重复记录：

select 
    u.employeeid [id_empleado], a.LOGDATETIME [checada], ASSETID [dispositivo]
from 
    [dbo].[AccessLog] a
inner join 
    [dbo].[BiometricDevice] b on (b.ID = a.MORPHOACCESSID)
inner join 
    [dbo].[User_] u on (a.userid = u.id)
order by 
    id_empleado, checada asc

输出示例：

id_empleado checada dispositivo
14907   2016-08-22 13:25:40.000 02
14907   2016-08-22 23:32:10.000 01
14907   2016-08-23 12:08:03.000 01
14907   2016-08-23 22:51:39.000 02
14907   2016-08-24 12:29:07.000 01
14907   2016-08-26 23:24:55.000 01
14907   2016-08-30 00:29:08.000 01
16034   2016-10-05 12:51:51.000 GERENCIA
16034   2016-10-06 04:03:56.000 GERENCIA
16034   2016-10-08 11:37:29.000 GERENCIA
16034   2016-10-09 10:52:51.000 GERENCIA
16034   2016-10-10 00:51:16.000 GERENCIA
16034   2016-10-10 11:19:42.000 GERENCIA

我需要什么：

id_empleado checada dispositivo
14907   2016-08-22 13:25:40.000 02
16034   2016-10-10 11:19:42.000 GERENCIA

Answer 1

请尝试以下方法。它会选择与给定checada相关联的第一个dispositivo和id_empleado。

select [id_empleado], [checada], [dispositivo]
from (
    select u.employeeid [id_empleado],a.LOGDATETIME [checada], ASSETID [dispositivo],
    ROW_NUMBER() OVER(PARTITION BY u.employeeid ORDER BY a.LOGDATETIME) as seq
    from [dbo].[AccessLog] a
    inner join [dbo].[BiometricDevice] b on (b.ID = a.MORPHOACCESSID)
    inner join [dbo].[User_] u on (a.userid = u.id)) as a
where seq = 1
order by id_empleado, checada asc

输出：

id_empleado checada                 dispositivo
----------- ----------------------- ------------------------------
14907       2016-08-22 13:25:40.000 02
16034       2016-10-05 12:51:51.000 GERENCIA

参考： MSDN ROW_NUMBER (Transact-SQL)

Answer 2

这应该有效：

select t.* 
from 
    (select 
         u.employeeid, a.logdatetime, assetid,
         row_number() over (partition by u.employeeid order by u.employeeid) as r   
     from 
         [dbo].[AccessLog] a  
     inner join 
         [dbo].[BiometricDevice] b on (b.ID = a.MORPHOACCESSID)   
     inner join 
         [dbo].[User_] u on (a.userid = u.id)   
) t  
where r = 1

Answer 3

使用上面帖子中提到的 GROUP BY 或 ROW_NUMBER（）

select 
    u.employeeid [id_empleado], MAX(a.LOGDATETIME) [checada], MAX(ASSETID) [dispositivo]
from 
    [dbo].[AccessLog] a
inner join 
    [dbo].[BiometricDevice] b on (b.ID = a.MORPHOACCESSID)
inner join 
    [dbo].[User_] u on (a.userid = u.id)
GROUP BY u.employeeid
order by 
    id_empleado, checada asc

避免重复的条目

3 个答案: