Sql 如何根据分组列A获取列B的值

我使用Microsoft SQL Server 2019。从一个表中，我想得到B列的值，这取决于a列的值是某个where子句上的最大值

让我们举个小例子

创建表：

CREATE TABLE [tblDummy]
(
    [ID] [int] NOT NULL,
    [SubID] [int] NULL,
    [PointInTime] [datetime] NULL,
    [Value] [decimal](18, 6) NULL,

    CONSTRAINT [PK_tblDummy] 
        PRIMARY KEY CLUSTERED ([ID] ASC)
                    WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, 
                          IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, 
                          ALLOW_PAGE_LOCKS = ON, OPTIMIZE_FOR_SEQUENTIAL_KEY = OFF) ON [PRIMARY]
) ON [PRIMARY]

INSERT INTO tblDummy (ID, SubID, PointInTime, Value) 
VALUES (1, 1, {d '2020-01-01'}, 1.1),
       (2, 2, {d '2020-02-01'}, 1.2),
       (3, 3, {d '2020-03-01'}, 1.3),
       (4, 1, {d '2020-04-01'}, 2.1),
       (5, 2, {d '2020-05-01'}, 2.2),
       (6, 3, {d '2020-06-01'}, 2.3),
       (7, 1, {d '2020-07-01'}, 3.1),
       (8, 2, {d '2020-08-01'}, 3.2),
       (9, 3, {d '2020-09-01'}, 3.3),
       (10, 1, {d '2020-10-01'}, 4.1),
       (11, 2, {d '2020-11-01'}, 4.2),
       (12, 3, {d '2020-12-01'}, 4.3)

将数据插入表中：

CREATE TABLE [tblDummy]
(
    [ID] [int] NOT NULL,
    [SubID] [int] NULL,
    [PointInTime] [datetime] NULL,
    [Value] [decimal](18, 6) NULL,

    CONSTRAINT [PK_tblDummy] 
        PRIMARY KEY CLUSTERED ([ID] ASC)
                    WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, 
                          IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, 
                          ALLOW_PAGE_LOCKS = ON, OPTIMIZE_FOR_SEQUENTIAL_KEY = OFF) ON [PRIMARY]
) ON [PRIMARY]

INSERT INTO tblDummy (ID, SubID, PointInTime, Value) 
VALUES (1, 1, {d '2020-01-01'}, 1.1),
       (2, 2, {d '2020-02-01'}, 1.2),
       (3, 3, {d '2020-03-01'}, 1.3),
       (4, 1, {d '2020-04-01'}, 2.1),
       (5, 2, {d '2020-05-01'}, 2.2),
       (6, 3, {d '2020-06-01'}, 2.3),
       (7, 1, {d '2020-07-01'}, 3.1),
       (8, 2, {d '2020-08-01'}, 3.2),
       (9, 3, {d '2020-09-01'}, 3.3),
       (10, 1, {d '2020-10-01'}, 4.1),
       (11, 2, {d '2020-11-01'}, 4.2),
       (12, 3, {d '2020-12-01'}, 4.3)

我想获得每个

SubID的
值，其中
PointInTime
是最大值（超过
SubID
）

我现在能想到的唯一方法是自连接，类似这样：

SELECT A.SubID, A.Value 
FROM tblDummy AS A
INNER JOIN 
    (SELECT SubID, MAX(PointInTime) AS PIT 
     FROM tblDummy 
     GROUP BY SubID) AS B ON A.SubID = B.SubID AND A.PointInTime = B.PIT

但我不喜欢子查询，也不喜欢联接使用非唯一的列。因此，如果数据包含两次“PointInTime”值（使用相同的“SubID”），那么由于连接，数据将加倍

有没有其他方法可以达到我的目标？也许使用
分区。。。语法问题？
我想您需要窗口功能：

select d.*
from (select d.*,
             row_number() over (partition by subid order by PointInTime desc) as seqnum
      from tblDummy d
     ) d
where seqnum = 1;

您还可以使用相关子查询来表示：

select d.*
from tblDummy d
where d.PointInTime = (select max(d2.PointInTime)
                       from tblDummy d2
                       where d2.subid = d.subid
                      );

这两个查询都可以受益于
tblDummy（subid，PointInTime）
上的索引，但这仍然是一个子查询。对于大量数据（比如一百万行），它是如何“运行”的？@Nostromo窗口函数更适合于大量行