当前位置: 首页 > 图文教程 > 数据库 > MSSQL > 一条语句简单解决“每个Y的最新X”的经典sql语句

MSSQL
获取MSSQL数据字典的SQL语句
Microsoft Search 服务无法启动 解决办法.
找出所有非xml索引并重新整理的sql
SQL Server提示"选定的用户拥有对象,所以无法除去该用户”
SQL Server"错误 21002: [SQL-DMO]用户 * 已经存在问题解决
删除重复记录，并且剩下一条
select * from sp_who的解决方案
Sql语句与存储过程查询数据的性能测试实现代码
MSSQL 游标使用 心得
SQL 时间类型的模糊查询
mssql 指定字段编号sql语句
sql server中的decimal或者numeric的精度问题
sql 数据库还原图文教程
存储过程解密(破解函数,过程,触发器,视图.仅限于SQLSERVER2000)
MSSQL SERVER中的BETWEEN AND的使用
SQL Server约束增强的两点建议
SQL Server 更改DB的Collation
Oracle 10g各个帐号的访问权限、登录路径、监控状态命令查询等等
SQL Server的复制功能
sql 普通行列转换

No. « 12 13 14 15 »

技术文章搜索

关键字

MSSQL 中的 一条语句简单解决“每个Y的最新X”的经典sql语句

出处:互联网   整理: 软晨网（RuanChen.com）   发布: 2009-09-13   浏览: 79 ::
收藏到网摘: n/a

“每个Y的最新X”是一个经典的SQL问题，工作中经常碰到。当然不是“按Y分组求最新的X值”那么简单，要求最新X的那条记录或主键ID。用一条SQL语句可以简单的解决此问题。 生成实例表和数据：
复制代码 代码如下:

/****** 创建表 ******/
if exists (select * from dbo.sysobjects where id = object_id(N'[dbo].[Table]') and OBJECTPROPERTY(id, N'IsUserTable') = 1)
drop table [dbo].[Table]
GO
Create TABLE [dbo].[Table] (
[ID] [int] IDENTITY (1, 1) NOT NULL ,
[Y] [nvarchar] (50) COLLATE Chinese_PRC_CI_AS NOT NULL ,
[X] [smalldatetime] NOT NULL
) ON [PRIMARY]
GO
--插入数据
Insert INTO [Table](Y, X) values('CCC', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('AAA', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('AAA', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('BBB', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('BBB', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('BBB', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('CCC', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('AAA', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('CCC', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('DDD', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('DDD', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('DDD', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('EEE', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('EEE', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('EEE', '2007-03-03 03:03:03')
GO
/****** 创建表 ******/
if exists (select * from dbo.sysobjects where id = object_id(N'[dbo].[Table]') and OBJECTPROPERTY(id, N'IsUserTable') = 1)
drop table [dbo].[Table]
GO
Create TABLE [dbo].[Table] (
[ID] [int] IDENTITY (1, 1) NOT NULL ,
[Y] [nvarchar] (50) COLLATE Chinese_PRC_CI_AS NOT NULL ,
[X] [smalldatetime] NOT NULL
) ON [PRIMARY]
GO
--插入数据
Insert INTO [Table](Y, X) values('CCC', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('AAA', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('AAA', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('BBB', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('BBB', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('BBB', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('CCC', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('AAA', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('CCC', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('DDD', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('DDD', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('DDD', '2007-03-03 03:03:03')
Insert INTO [Table](Y, X) values('EEE', '2007-01-01 01:01:01')
Insert INTO [Table](Y, X) values('EEE', '2007-02-02 02:02:02')
Insert INTO [Table](Y, X) values('EEE', '2007-03-03 03:03:03')

GO解决“每个Y的最新X”经典SQL问题：以下几种方法真是八仙过海
复制代码 代码如下:

Select ID, Y, X
FROM [Table] T1
Where (NOT EXISTS
(Select 1
FROM [Table] T2
Where (T2.Y = T1 .Y) AND (T2.X > T1 .X or
T2.X = T1 .X AND T2.ID > T1 .ID)))
/*****************************************************************************/
Select *
FROM [Table]
Where ID IN
(Select MAX(T1.ID)
FROM [Table] T1 JOIN
(Select y, MAX(x) x
FROM [Table]
GROUP BY y) T2 ON T1.Y = T2.Y AND T1.x = T2.x
GROUP BY T1.y)
/*****************************************************************************/
Select T .ID, T .Y, T .X
FROM [Table] T INNER JOIN
(Select MAX(T1.ID) AS ID
FROM [Table] T1 JOIN
(Select y, MAX(x) x
FROM [Table]
GROUP BY y) T2 ON T1.Y = T2.Y AND T1.x = T2.x
GROUP BY T1.y) T2 ON T .ID = T2.ID
/*****************************************************************************/
Select *
FROM [Table] T1
Where ID IN
(Select TOP 1 ID
FROM [Table]
Where Y = T1.Y
orDER BY X DESC)
/*****************************************************************************/
Select *
FROM [Table] T1
Where (ID =
(Select TOP 1 ID
FROM [Table]
Where Y = T1.Y
orDER BY X DESC, ID DESC))
/*****************************************************************************/
/*****************************************************************************/
Select ID, Y, X
FROM [Table] T1
Where (NOT EXISTS
(Select 1
FROM [Table] T2
Where (T2.Y = T1 .Y) AND (T2.X > T1 .X or
T2.X = T1 .X AND T2.ID > T1 .ID)))
/*****************************************************************************/
Select *
FROM [Table]
Where ID IN
(Select MAX(T1.ID)
FROM [Table] T1 JOIN
(Select y, MAX(x) x
FROM [Table]
GROUP BY y) T2 ON T1.Y = T2.Y AND T1.x = T2.x
GROUP BY T1.y)
/*****************************************************************************/
Select T .ID, T .Y, T .X
FROM [Table] T INNER JOIN
(Select MAX(T1.ID) AS ID
FROM [Table] T1 JOIN
(Select y, MAX(x) x
FROM [Table]
GROUP BY y) T2 ON T1.Y = T2.Y AND T1.x = T2.x
GROUP BY T1.y) T2 ON T .ID = T2.ID
/*****************************************************************************/
Select *
FROM [Table] T1
Where ID IN
(Select TOP 1 ID
FROM [Table]
Where Y = T1.Y
orDER BY X DESC)
/*****************************************************************************/
Select *
FROM [Table] T1
Where (ID =
(Select TOP 1 ID
FROM [Table]
Where Y = T1.Y
orDER BY X DESC, ID DESC))

/*****************************************************************************
/效率嘛,在不同的字段建立索引速度都不尽相同,使用者见仁见智了.
第一种方法速度在各方面都不错,而且在Y列在建立索引，可以大大优化查询速度。