Question

我在表格中的基本选择中的这个片段是从3秒到8进行查询。任何想法？

case 
   when @excludeprojtag = 1 
      then qtyruletag 
      else 0 
end as qtyruletag

编辑：这里是整个查询

select case when @excludeprojtag = 1 then qtyruletag else 0 end as qtyruletag,listid,quantity
from tb_sales where date between '1/1/2015' and '1/1/2016' 
AND CASE WHEN @excludeTestOrders = 1 THEN AccountID ELSE 123 END <> 1234

Answer 1

USE AdventureWorks -- 2012 version
GO       
DECLARE @includezip INT = 0

SELECT @includezip = DATEPART(ss,GETDATE()) % 2
SET STATISTICS TIME ON;

SELECT pa.AddressLine1, pa.City,
    CASE -- CASE method
        WHEN @includezip = 1
            THEN pa.PostalCode
            ELSE ''
        END AS ZipCode
FROM Person.BusinessEntity pbe
    LEFT JOIN Person.BusinessEntityAddress pbea
        ON pbe.BusinessEntityID = pbea.BusinessEntityID
    LEFT JOIN Person.Address pa
        ON pbea.AddressID = pa.AddressID

SELECT pa.AddressLine1, pa.City,
    SUBSTRING(pa.PostalCode,1,LEN(pa.PostalCode) * @includezip) AS ZipCode -- No CASE used
FROM Person.BusinessEntity pbe
    LEFT JOIN Person.BusinessEntityAddress pbea
        ON pbe.BusinessEntityID = pbea.BusinessEntityID
    LEFT JOIN Person.Address pa
        ON pbea.AddressID = pa.AddressID
GO
SET STATISTICS TIME OFF;
GO

（受影响的20812行）

SQL Server执行时间： CPU时间= 15 ms，经过时间= 102 ms。

（受影响的20812行）

SQL Server执行时间： CPU时间= 16 ms，经过时间= 140 ms。

我经常跑了很多次。在我看来，第二个总是比较慢。当我第一次创建这个时，我认为zip是数字的，但是我忘了加拿大等等。所以如果由于某种原因你反对CASE语句的一般原则，你可以在有二进制参数时使用这个方法。在这里，我只是将其设置为或多或少随机。

Answer 2

我认为哪里可以改进

select case when @excludeprojtag = 1 then qtyruletag else 0 end as qtyruletag
     , listid, quantity
from tb_sales 
where date between '1/1/2015' and '1/1/2016' 
  AND ( @excludeTestOrders <> 1 OR AccountID <> 1234 )

tsql - 在select子句中优化case语句

2 个答案: