Question

我在SQL Server 2005上运行了一个T-SQL查询，我在其中定义了一个要返回给应用程序的数据表。

我使用几个case语句分配几个值，每个case语句都有相同的逻辑。

我想知道是否有办法让这个效率更高，而不是多次执行相同的逻辑？我在这里缩写了逻辑，但希望有足够的说明。有没有办法让我不必使用相同的“pt.PTYP_Category IN（90,92,94,96,98）”逻辑来定义三个case语句？

`INSERT INTO @MemberAccountingDetail
    (
     MACD_DPST_FEES_RowID,
     MACD_HighOrderSort,
     MACD_CategorySortDate, 
     MACD_MemberName
     )  
 SELECT 
   d.DPST_ID As MACD_DPST_FEES_RowID,
   CASE
         WHEN  pt.PTYP_Category IN (90, 92, 94, 96, 98) 
             THEN 0
         ELSE     1
     END As MACD_HighOrderSort,
   CASE
         WHEN  pt.PTYP_Category IN (90, 92, 94, 96, 98) 
             THEN DATEADD(ss, 1, d.DPST_TransactionDate)
         ELSE     @DefaultDate
     END As MACD_CategorySortDate,
   CASE
         WHEN  pt.PTYP_Category IN (90, 92, 94, 96, 98) 
             THEN pt.PTYP_CategoryDesc
         ELSE     SUBSTRING(ISNULL(d.DPST_OriginUserID, ''), 1, 11)
     END As MACD_MemberName
 FROM
     Deposit d
        INNER JOIN PaymentType pt
           ON pt.PTYP_ID = d.DPST_PTYP_ID
 WHERE 
     d.DPST_MEMB_ID = @MEMB_ID;

谢谢你，吉姆`

Answer 1

您可以使用公用表表达式，因此您只需编写一次“IN STATEMENT”：

INSERT INTO @MemberAccountingDetail
(
    MACD_DPST_FEES_RowID,
    MACD_HighOrderSort,
    MACD_CategorySortDate, 
    MACD_MemberName
)
WITH CTE (PTYP_ID, IsInCollection)
AS
(
    SELECT pt.PTYP_ID, CASE WHEN  pt.PTYP_Category IN (90, 92, 94, 96, 98) THEN 0 ELSE 1 END AS IsInCollection
    FROM PaymentType pt
)
SELECT 
    d.DPST_ID As MACD_DPST_FEES_RowID,
    CTE.IsInCollection As MACD_HighOrderSort,
    CASE WHEN CTE.IsInCollection = 1 
        THEN DATEADD(ss, 1, d.DPST_TransactionDate)
        ELSE     @DefaultDate
    END As MACD_CategorySortDate,
    CASE WHEN CTE.IsInCollection = 1
        THEN pt.PTYP_CategoryDesc
        ELSE SUBSTRING(ISNULL(d.DPST_OriginUserID, ''), 1, 11)
    END As MACD_MemberName
FROM Deposit d
INNER JOIN CTE 
  ON CTE.PTYP_ID = d.DPST_PTYP_ID
INNER JOIN PaymentType pt
  ON pt.PTYP_ID = d.DPST_PTYP_ID
WHERE d.DPST_MEMB_ID = @MEMB_ID;

Answer 2

您可以尝试使用UNION进行两次单独选择，但我不确定它会更快：

INSERT INTO @MemberAccountingDetail
    (
     MACD_DPST_FEES_RowID,
     MACD_HighOrderSort,
     MACD_CategorySortDate, 
     MACD_MemberName
     )  

 SELECT 
   d.DPST_ID As MACD_DPST_FEES_RowID,
   0 As MACD_HighOrderSort,
   DATEADD(ss, 1, d.DPST_TransactionDate) As MACD_CategorySortDate,
   pt.PTYP_CategoryDesc As MACD_MemberName
 FROM
     Deposit d
        INNER JOIN PaymentType pt
           ON pt.PTYP_ID = d.DPST_PTYP_ID
 WHERE 
     d.DPST_MEMB_ID = @MEMB_ID and pt.PTYP_Category IN (90, 92, 94, 96, 98)
 UNION ALL
 SELECT 
   d.DPST_ID As MACD_DPST_FEES_RowID,
   1 As MACD_HighOrderSort,
   @DefaultDate As MACD_CategorySortDate,
   SUBSTRING(ISNULL(d.DPST_OriginUserID, ''), 1, 11) as MACD_MemberName
 FROM
     Deposit d
        INNER JOIN PaymentType pt
           ON pt.PTYP_ID = d.DPST_PTYP_ID
 WHERE 
     d.DPST_MEMB_ID = @MEMB_ID and pt.PTYP_Category NOT IN (90, 92, 94, 96, 98)

Answer 3

您可以在表格中创建一个计算列，将这5个类别汇总到一个代码中，然后使用它。

这意味着您不再需要硬编码魔术数字。但是你仍然需要四种不同的案例陈述。这是任何试图解决它的“聪明”的事情之一，只会使它变得更加复杂。

Answer 4

CROSS APPLY的另一个选项

 INSERT INTO @MemberAccountingDetail
    (
     MACD_DPST_FEES_RowID,
     MACD_HighOrderSort,
     MACD_CategorySortDate, 
     MACD_MemberName
     )  
 SELECT 
   d.DPST_ID As MACD_DPST_FEES_RowID,
   CASE WHEN o.IsMatch = 1 THEN 0 
        ELSE 1 END As MACD_HighOrderSort,
   CASE WHEN o.IsMatch = 1 THEN DATEADD(ss, 1, d.DPST_TransactionDate) 
        ELSE @DefaultDate END As MACD_CategorySortDate,
   CASE WHEN o.IsMatch = 1 THEN pt.PTYP_CategoryDesc 
        ELSE SUBSTRING(ISNULL(d.DPST_OriginUserID, ''), 1, 11) END As MACD_MemberName
 FROM Deposit d INNER JOIN PaymentType pt ON pt.PTYP_ID = d.DPST_PTYP_ID
                CROSS APPLY (SELECT CASE WHEN pt.PTYP_Category IN (90, 92, 94, 96, 98) THEN 1 END AS IsMatch) o
 WHERE d.DPST_MEMB_ID = @MEMB_ID;

如何将SQL Server查询逻辑整合到一个案例中？

4 个答案: