SQL可以由可能的三個值或者根本沒有

Question

拉記錄假設你有WIDGETS表：

• 爲widgetid INT
• ColorID INT（查找值：紅，藍，綠，黃，黑，布朗）
• SizeID INT（查找值：小，中，大，大）
• 重量INT（查找值：超輕，輕，正常，重，UltraHeavy

好吧，這是表的一般想法。我不需要查找名稱，我所擁有的是lookupvalue ID。所以從這個表，我就能夠拉列表回到符合下面要說的crieria：

1. 告訴我誰是紅色或藍色或黑色的所有部件
2. 告訴我誰是紅色或所有小工具藍，和小或醫學
3. 告訴我誰是沉重的所有部件只
4. 告訴我誰是沉重和黃所有小
5. 告訴我誰是沉重和黃大或小的所有部件
6. 告訴我所有隻有大號的小配件只有
7. 讓我看看所有隻有綠色的小工具

想到這個主意嗎？我一直在嘗試使用Stored Proc，它允許我發送某種類型的參數。即使嘗試了動態SQL，但得到奇怪的錯誤。現在不記得了。

我嘗試3，排序的作品，如果我能得到它在一個Proc工作，並能夠調節出其加入的我想要什麼，我已經試過

例子：

注：以我的例子：（CourseID是爲widgetid）和（STATEID是ColorID）和（CreditTypeID是SizeID）和（SubjectID是WeightID）

嘗試方法1

ALTER PROCEDURE [dbo].[CourseListFullInfoByStateCreditSubject] 
    @StateIDs VARCHAR(200) = '', 
    @CreditTypeIDs VARCHAR(200) = '', 
    @SubjectTypeIDs VARCHAR(200) = '' 
AS 

BEGIN 

    DECLARE @SQL AS NVARCHAR(MAX) 
    SET @SQL = 'SELECT DISTINCT 
        C.CourseID, 
        LU.FirstName, 
        LU.LastName, 
        (SELECT COUNT(ReviewID) FROM Review AS R WHERE R.CourseID = C.CourseID) AS ReviewCount 
       FROM [Course] AS C WITH(NOLOCK) 
       JOIN LexUser AS LU ON LU.LexUserID = C.PresenterID ' 

If @StateIDs IS NOT NULL AND @StateIDs <> '''' AND @StateIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + ' JOIN CourseToState AS CS ON CS.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CS.StateID AS VARCHAR) + '','', '','' + @StateIDs + '','') > 0 ' 
END 

If @CreditTypeIDs IS NOT NULL AND @CreditTypeIDs <> '' AND @CreditTypeIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + 'JOIN CourseToCreditType As CC ON CC.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CC.CreditTypeID AS VARCHAR) + '','', '','' + @CreditTypeIDs + '','') > 0 ' 
END 

If @SubjectTypeIDs IS NOT NULL AND @SubjectTypeIDs <> '' AND @SubjectTypeIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + 'JOIN CourseToSubject As CSu ON CSu.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CSu.SubjectID AS VARCHAR) + '','', '','' + @SubjectTypeIDs + '','') > 0 ' 
END 

EXEC sp_executesql @SQL 

有了嘗試1我嘗試發送的ID的：

[CourseListFullInfoByStateCreditSubject] ''1,2,4'', ''0'', ''0'' 

...但我得到一個錯誤，「附近有語法錯誤 '1'。「

嘗試方法2，使同樣的錯誤

DECLARE @SQL AS NVARCHAR(MAX) 
SET @SQL = 'SELECT DISTINCT 
        C.CourseID, 
        LU.FirstName, 
        LU.LastName, 
        (SELECT COUNT(ReviewID) 
        FROM Review AS R 
        WHERE R.CourseID = C.CourseID) AS ReviewCount 
      FROM [Course] AS C WITH(NOLOCK) 
      JOIN LexUser AS LU ON LU.LexUserID = C.PresenterID ' 

If @StateIDs IS NOT NULL AND @StateIDs <> '''' AND @StateIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + ' JOIN CourseToState AS CS ON CS.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CS.StateID AS VARCHAR) + '','', '','' + @StateIDs + '','') > 0 ' 
END 

If @CreditTypeIDs IS NOT NULL AND @CreditTypeIDs <> '' AND @CreditTypeIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + 'JOIN CourseToCreditType As CC ON CC.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CC.CreditTypeID AS VARCHAR) + '','', '','' + @CreditTypeIDs + '','') > 0' 
END 

If @SubjectTypeIDs IS NOT NULL AND @SubjectTypeIDs <> '' AND @SubjectTypeIDs <> '0' 
BEGIN 
    SET @SQL = @SQL + 'JOIN CourseToSubject As CSu ON CSu.CourseID = C.CourseID AND CHARINDEX('','' + CAST(CSu.SubjectID AS VARCHAR) + '','', '','' + @SubjectTypeIDs + '','') > 0' 
END 

嘗試方法3 - 排序作品 此方法僅如果我把它的ID對每個菲爾德的作品，我不能離開一個空白得到「所有」，我不能把它在一個存儲過程，而這正是我試圖嘗試＃做2

DECLARE @StateIDs VARCHAR(200) = '' 
DECLARE @CreditTypeIDs VARCHAR(200) = '' 
DECLARE @SubjectTypeIDs VARCHAR(200) = '' 
SET @StateIDs = '1,3,2,' 
SET @CreditTypeIDs = '1,3' 
SET @SubjectTypeIDs = '1,2,3,4' 

SELECT DISTINCT 
     C.CourseID, 
     LU.FirstName, 
     LU.LastName, 
     (SELECT COUNT(ReviewID) FROM Review AS R WHERE R.CourseID = C.CourseID) AS ReviewCount 
    FROM [Course] AS C WITH(NOLOCK) 
    JOIN LexUser AS LU ON LU.LexUserID = C.PresenterID 
    JOIN CourseToState AS CS ON CS.CourseID = C.CourseID AND CHARINDEX(',' + CAST(CS.StateID AS VARCHAR) + ',', ',' + @StateIDs + ',') > 0 
    JOIN CourseToCreditType As CC ON CC.CourseID = C.CourseID AND CHARINDEX(',' + CAST(CC.CreditTypeID AS VARCHAR) + ',', ',' + @CreditTypeIDs + ',') > 0 
    JOIN CourseToSubject As CSu ON CSu.CourseID = C.CourseID AND CHARINDEX(',' + CAST(CSu.SubjectID AS VARCHAR) + ',', ',' + @SubjectTypeIDs + ',') > 0 

Answer 1

怎麼是這樣的：

請用逗號分隔的條件列表中的存儲過程，應該像

DECLARE @colorConditions VARCHAR(100) 
SET @colorConditions = ' Red, Blue, Green, Yellow, Black, Brown, ' 

DECLARE @sizeConditions VARCHAR(100) 
SET @sizeConditions = ' Small, Med, Big, Large, ' 

DECLARE @weightConditions VARCHAR(100) 
SET @weightConditions = ' UltraLight, Light, Normal, Heavy, UltraHeavy, ' 

DECLARE @SQL VARCHAR(1000) 
SET @SQL = 
    'SELECT * 
     FROM Widgets W 
     INNER JOIN Colors C 
     ON W.ColorID = C.ColorID AND ''' + @colorConditions + ''' LIKE ''% '' + C.Color + '', %'' 
     INNER JOIN Sizes S 
     ON W.SizeID = S.SizeID AND ''' + @sizeConditions + ''' LIKE ''% '' + S.Size + '', %'' 
     INNER JOIN Weight WT 
     ON W.WeightID = W.WeightID AND ''' + @weightConditions + ''' LIKE '' %'' + WT.Weight + '', %''' 

EXEC (@SQL) 

這裏沒有什麼特別。它所做的只是將每個特徵（顏色，大小，重量）與符合條件的逗號分隔列表進行匹配。如果你想過濾藍色項目，你應該傳遞字符串'藍色'。側面的空格和逗號應該存在，因爲它們更容易在那裏與SQL進行對抗。然後，條件字符串比較像這樣每種顏色值：

Condition string: 

' Blue, Green, ' 

Colors: 

Red 
Blue 
Green 

...AND ' Blue, Green, ' LIKE '% Red, %' -- fails 
...AND ' Blue, Green, ' LIKE '% Blue, %' -- succeeds 
...AND ' Blue, Green, ' LIKE '% Green, %' -- succeeds 

這必須動態執行，除非你分割開出的條件不知何故到一個臨時表，並使用IN操作。

Answer 2

嘗試從左邊小工具加入到所有的查找工作並使用Where過濾出相應的行。

Answer 3

試試這個......來調用它，對於每個參數，傳遞null來忽略，或者位掩碼 你想包括的值的組合。根據以下圖表 ...

-- ----Color ---- 
Red = 1 
Blue = 2 
Green = 4 
Yellow = 8 
Black = 16 
Brown = 32 
Red Or Green = 5 

-- ---- Size --- 
Small = 1 
Medium = 2 
Big = 4 
Large = 8 

-- ----Weight ---- 
UltraLight = 1 
Light  = 2 
Normal  = 4 
Heavy  = 8 
UltraHeavy = 16 

此外，改變在數據庫（在查找表中兩者的PK在窗口小部件表，FKS）查找的整數值，使得它們的所有的兩個適當的權力（1,2,4，8,16 ...）

-- ------------------------- 
Create procedure GetWidgets 
@Color TinyInt = Null, 
@Size TinyInt = Null 
@Weight TinyInt = Null 
As 
Set NoCount On 

-- ------------------------------- 

    Select * From Widgets 
    Where Color & IsNull(@Color, Color)> 0 
     And Size & IsNull (@Size,Size) > 0 
     And Weight & IsNull (@Weight, Weight) > 0 

    Return 0 
-- ------------------------- 

---

Answer 4

此解決方案的前提條件是分裂表值函數。網上有很多例子，但我建議使用使用Tally或「Numbers」表的變體，因爲它是最快的。一旦你有了，解決方案有點微不足道：

Declare @StateIdList nvarchar(max) 
Declare @CreditTypeIdList nvarchar(max) 
Declare @SubjectTypeIds nvarchar(max) 

Select F1....Fn 
From Course As C 
Where (
     @StateIdList Is Null 
     Or Exists(
        Select 1 
        From CourseToState As CS1 
         Join dbo.Split(@StateIdList) As S1 
          On S1.Id = CS1.Id 
        Where CS1.CourseId = C.CourseId 
        ) 
     ) 
    And (
     @CreditTypeIdList Is Null 
     Or Exists(
        Select 1 
        From CourseToCreditType As CT1 
         Join dbo.Split(@CreditTypeIdList) As S1 
          On S1.Id = CT1.Id 
        Where CT1.CourseId = C.CourseId 
        ) 
     ) 
    And (
     @SubjectTypeIds Is Null 
     Or Exists(
        Select 1 
        From CourseToSubject As CSu1 
         Join dbo.Split(@CreditTypeIdList) As S1 
          On S1.Id = CSu1.Id 
        Where CSu1.CourseId = C.CourseId 
        ) 
     )