【转】在一个SQL Server表中一行的多个列找出最大值

时间 2019-11-25

原文原文链接

https://www.cnblogs.com/lyhabc/archive/2015/10/21/4899244.htmlhtml

有时候咱们须要从多个相同的列里（这些列的数据类型相同）找出最大的那个值，并显示spa

这里给出一个例子3d

IF (OBJECT_ID('tempdb..##TestTable') IS NOT NULL)
    DROP TABLE ##TestTable

CREATE TABLE ##TestTable
(
    ID INT IDENTITY(1,1) PRIMARY KEY,
    Name NVARCHAR(40),
    UpdateByApp1Date DATETIME,
    UpdateByApp2Date DATETIME,
    UpdateByApp3Date DATETIME

)

INSERT INTO ##TestTable(Name, UpdateByApp1Date, UpdateByApp2Date, UpdateByApp3Date )
VALUES('ABC', '2015-08-05','2015-08-04', '2015-08-06'),
      ('NewCopmany', '2014-07-05','2012-12-09', '2015-08-14'),
      ('MyCompany', '2015-03-05','2015-01-14', '2015-07-26')
      
SELECT * FROM ##TestTable

结果以下所示code

有三种方法能够实现htm

方法一blog

SELECT  ID ,
        Name ,
        ( SELECT    MAX(LastUpdateDate)
          FROM      ( VALUES ( UpdateByApp1Date), ( UpdateByApp2Date),
                    ( UpdateByApp3Date) ) AS UpdateDate ( LastUpdateDate )
          ) AS LastUpdateDate
FROM    ##TestTable

方法二get

SELECT  ID ,
        [Name] ,
        MAX(UpdateDate) AS LastUpdateDate
FROM    ##TestTable UNPIVOT ( UpdateDate FOR DateVal IN ( UpdateByApp1Date,
                                                          UpdateByApp2Date,
                                                          UpdateByApp3Date ) ) AS u
GROUP BY ID ,
        Name

方法三io

SELECT  ID ,
        name ,
        ( SELECT    MAX(UpdateDate) AS LastUpdateDate
          FROM      ( SELECT    tt.UpdateByApp1Date AS UpdateDate
                      UNION
                      SELECT    tt.UpdateByApp2Date
                      UNION
                      SELECT    tt.UpdateByApp3Date
                    ) ud
        ) LastUpdateDate
FROM    ##TestTable tt

第一种方法使用values子句，将每行数据构造为只有一个字段的表，之后求最大值，很是巧妙ast

第二种方法使用行转列常常用的UNPIVOT 关键字进行转换再显示class

第三种方法跟第一种方法差很少，可是使用union将三个UpdateByAppDate字段合并为只有一个字段的结果集而后求最大值

第一种方法的执行计划

第二种方法的执行计划

第三种方法的执行计划

总的来讲，第一种方法的执行计划是最好的

注意，这里不涉及分组

IF (OBJECT_ID('tempdb..##TestTable') IS NOT NULL)
    DROP TABLE ##TestTable

CREATE TABLE ##TestTable
    (
      ID INT IDENTITY(1, 1)
             PRIMARY KEY ,
      Name NVARCHAR(40) ,
      UpdateByApp1Date DATETIME ,
      UpdateByApp2Date DATETIME ,
      UpdateByApp3Date DATETIME
    )

INSERT  INTO ##TestTable
        ( Name, UpdateByApp1Date, UpdateByApp2Date, UpdateByApp3Date )
VALUES  ( 'ABC', '2015-08-05', '2015-08-04', '2015-08-06' ),
        ( 'ABC', '2015-07-05', '2015-06-04', '2015-09-06' ),
        ( 'NewCopmany', '2014-07-05', '2012-12-09', '2015-08-14' ),
        ( 'MyCompany', '2015-03-05', '2015-01-14', '2015-07-26' )
      
SELECT  *
FROM    ##TestTable

SELECT  ID ,
        Name ,
        ( SELECT    MAX(LastUpdateDate)
          FROM      ( VALUES ( UpdateByApp1Date), ( UpdateByApp2Date),
                    ( UpdateByApp3Date) ) AS UpdateDate ( LastUpdateDate )
          ) AS LastUpdateDate
FROM    ##TestTable

name列相同的话，是没法得出name分组以后的最大值，这里要注意一下