选择在SQL Server中最大的日期和COMPRATE的总和

Question

我想获得的总和（COMPRATE）仅MAX（effdt），具有MAX（ID）从下列查询，上述查询的

SELECT id, employee_code, effdt, comprate 
FROM   dbo.employees_compensation 
WHERE  (employee_code = '000321514') 
ORDER BY effdt desc, id desc 

结果如下，

id  employee_code effdt     comprate 
1473433 000321514  2010-09-02 00:00:00.000 46000 
1473432 000321514  2010-09-02 00:00:00.000 138000 
1453627 000321514  2010-09-02 00:00:00.000 46000 
1453626 000321514  2010-09-02 00:00:00.000 138000 
1447165 000321514  2010-09-02 00:00:00.000 46000 
1447164 000321514  2010-09-02 00:00:00.000 138000 
1424948 000321514  2010-09-02 00:00:00.000 46000 
1424947 000321514  2010-09-02 00:00:00.000 138000 
1405789 000321514  2010-09-02 00:00:00.000 46000 
1405788 000321514  2010-09-02 00:00:00.000 138000 

我试过许多查询，但我没有得到预期的结果。我要的是从上面的结果，即顶前两个值，

id employee_code effdt     comprate 
1473433 000321514 2010-09-02 00:00:00.000 46000 
1473432 000321514 2010-09-02 00:00:00.000 138000 

，然后得到上述结果的总和（COMPRATE）即

sum(comprate) 
184000 

你能帮我得到预期的结果？提前致谢。

Answer 1

以下情况如何？

SELECT  SUM(comprate) AS ComprateSum 
FROM   (SELECT  TOP (2) id, employee_code, effdt, comprate 
         FROM   _Test 
         WHERE  (employee_code = '000321514') 
         ORDER BY effdt DESC, id DESC) AS derivedtbl_1 

Answer 2

我会建议查找如何在这里限制结果集的大小LIMIT 10..20 in SQL Server。然后，如果你使用SQL Server 2005 或新的（你没有在你的问题中指定），你可以使用一个CTE（公共表表达式）与如何使这里http://www.sqlteam.com/article/aggregating-correlated-sub-queries

Answer 3

子查询计算该ROW_NUMBER功能：

;WITH OrderedData AS 
(
    SELECT 
     id, employee_code, effdt, comprate, 
     RowNum = ROW_NUMBER() OVER(ORDER BY effdt DESC, id DESC) 
    FROM   
     dbo.employees_compensation 
    WHERE  
     employee_code = '000321514' 
    FROM dbo.YourTable 
) 
SELECT 
    SUM(comprate) 
FROM 
    OrderedData 
WHERE 
    RowNum <= 2 

这CTE订单数据和ROW_NUMBER功能手了序列号，从1开始，并通过effdt DESC, id DESC订购 - 所以最新的行（最新effdt和最高id）得到RowNum = 1。那么我基本上选择最上面的两行并且总结它们。