最大值（值）的 Select 列

Question

我有这些桌子

记录：

+----+-------------+--------------+-------------------------+---------+
| Id | IdHealtcare | ValueMonitor |       DateMonitor       | Checked |
+----+-------------+--------------+-------------------------+---------+
|  1 |           1 |           80 | 2020-05-14 19:40:11.873 | NULL    |
|  2 |           2 |           66 | 2020-05-14 20:56:33.360 | NULL    |
|  3 |           1 |          100 | 2020-05-14 20:32:24.397 | NULL    |
|  4 |           3 |           87 | 2020-05-14 20:57:12.390 | NULL    |
+----+-------------+--------------+-------------------------+---------+

卫生保健：

+----+-----------+----------+--------+------------+
| Id | IdPatient | IdDoctor | IdType | RangeValue |
+----+-----------+----------+--------+------------+
|  1 |         1 |        1 |      1 | 0-180      |
|  2 |         1 |        1 |      2 | 50-200     |
|  3 |         2 |        2 |      1 | 90-180     |
+----+-----------+----------+--------+------------+

我想为每种患者的医疗保健类型 select 提供患者的 ID 和最近的日期值。

我已经尝试过这些查询：

select p.CodiceFiscale, p.Surname, p.Name, t.Name, Max(r.DateMonitor), 
   r.ValueMonitor
from dbo.Patients as p
inner join dbo.HealtcareMonitoring as m on m.IdPatient = p.Id 
inner join dbo.Records as r on r.IdHealtcare = m.Id 
inner join dbo.MonitoringTypes as t on t.Id = m.IdType
where m.IdDoctor = 1
group by t.Name, p.CodiceFiscale, p.Surname, p.Name

SELECT m.IdPatient, m.IdType, MAX(r.DateMonitor), r.ValueMonitor AS LastDate
FROM  dbo.Records as r
inner join dbo.HealtcareMonitoring as m on m.Id = r.IdHealtcare  
where m.IdDoctor = 1
GROUP BY m.IdPatient, m.IdType

但我不能 select 该值，因为它不包含在聚合 function 或“分组依据”子句中。

我该如何解决？ 有没有更好的方法来解决这个问题？

Answer 1

你为什么不在group by中添加r.ValueMonitor ？ 它应该修复该错误

SELECT m.IdPatient, m.IdType, MAX(r.DateMonitor), r.ValueMonitor AS LastDate
    FROM  dbo.Records as r inner join dbo.HealtcareMonitoring as m
    on m.Id = r.IdHealtcare  
    where m.IdDoctor = 1
    GROUP BY m.IdPatient, m.IdType, r.ValueMonitor 

Answer 2

一种解决方案可以按非聚合列分组

select  p.CodiceFiscale, p.Surname, p.Name, t.Name, Max(r.DateMonitor), r.ValueMonitor from dbo.Patients as p
    inner join dbo.HealtcareMonitoring as m
    on m.IdPatient = p.Id 
    inner join dbo.Records as r
    on r.IdHealtcare = m.Id 
    inner join dbo.MonitoringTypes as t
    on t.Id = m.IdType
    where m.IdDoctor = 1 group by t.Name, p.CodiceFiscale, p.Surname, p.Name
group by p.CodiceFiscale, p.Surname, p.Name, t.Name, r.ValueMonitor

另一种解决方案可能是使用子查询，但性能更差

select  p.CodiceFiscale, p.Surname, p.Name, t.Name, 
(select MAX(DateMonitor) from dbo.Records where IdHealtcare = m.Id)
, r.ValueMonitor from dbo.Patients as p
    inner join dbo.HealtcareMonitoring as m
    on m.IdPatient = p.Id 
    inner join dbo.MonitoringTypes as t
    on t.Id = m.IdType
    where m.IdDoctor = 1 group by t.Name, p.CodiceFiscale, p.Surname, p.Name

更多关于加入与子查询的性能在这里SQL 加入与 SQL 子查询（性能）？