Question

我正在尝试使用COALESCE连接两行或三行，条件是EntityId在表中相等，以便我的输出如下：

EmailAddress                         EntityId

12_abc@abc.com::::13_pqr@pqr.com     101
12_koi@koi.com::::13_poi@poi.com     102

现在我根据我的查询获得以下结果：

EmailAddress   EmailTypeId     EntityId

abc@abc.com    12              101
pqr@pqr.com    13              101
koi@koi.com    12              102
poi@poi.com    13              102

我使用的查询如下：

DECLARE @MyTableVariable TABLE
(

  EmailAddress nvarchar(250),
  EntityIDE int
);

INSERT @MyTableVariable(
  EmailAddress,
  EntityIDE
 )

SELECT  concat([EmailTypeID], '_____', [EmailAddress]))), [EntityID]
FROM [dbo].[Email_Addresses] where [EntityID] in 

(select [EntityID] from [Entities] where [SourcePrimaryKey] in 
    (select [MerchantOwnerID] from [dbo].[Merchant_Owners] where [MerchantID] = 1 ) and
    [EntityTypeID] = (select [EntityTypeID] from [Entity_Types] where [EntityType] = 'MerchantOwner'))

select * from @MyTableVariable

我已经定义了一个表值参数，在我的select查询中，我将两列连成一列，但我不知道如何使用coalesce来连接EntityId相同的两行或更多行。

我where后的查询返回EntityId

列的列表

(select [EntityID] from [Entities] where [SourcePrimaryKey] in 
    (select [MerchantOwnerID] from [dbo].[Merchant_Owners] where [MerchantID] = 1 ) and
    [EntityTypeID] = (select [EntityTypeID] from [Entity_Types] where [EntityType] = 'MerchantOwner'))

以上查询返回以下结果集

EntityId

101
102
103
104
...

修改

; WITH CTE AS (
    SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID]
    FROM [dbo].[Email_Addresses] EA
    JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID])
    ,
    ABC
    As
    (
    SELECT EB.[PhoneNumber], EB.[EntityID], EB.[PhoneType]
    FROM [dbo].[Phones] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
    )
SELECT STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
            FROM CTE
            WHERE [EntityID] = T1.[EntityID]
            ORDER BY [EmailTypeID]
            FOR XML PATH ('')),1,4,'') [EmailAddress],
            T1.[EntityID],
            STUFF((SELECT '::::' + CONVERT(NVARCHAR, [PhoneType]) + '_' + [PhoneNumber]
            FROM ABC
            WHERE [EntityID] = T2.[EntityID]
            ORDER BY [PhoneType]
            FOR XML PATH ('')),1,4,'') [PhoneNumber],
            T2.[EntityID] 

FROM CTE T1, ABC T2
GROUP BY T1.[EntityID],T2.[EntityID]

这个结果给了我56行我的一个CTE返回8行，另一个CTE（ABC）返回7行

我不知道为什么它会成倍增加

EDIT2

; WITH CTE AS (
    SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID],
 row_number() over(partition by EA.EntityID order by EA.EntityID desc) as rn
    FROM [dbo].[Email_Addresses] EA
    JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
 ),
 ABC
 As
 (
 SELECT EB.[PhoneNumber], EB.[EntityID], EB.[PhoneType],
 row_number() over(partition by EB.EntityID order by EB.EntityID desc) as rn
    FROM [dbo].[Phones] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
 ),
 GHK
 As
 (
  SELECT EB.[Address1], EB.[City], EB.[State],EB.[EntityID],
  row_number() over(partition by EB.EntityID order by EB.EntityID desc) as rn
    FROM [dbo].[Addresses] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
 ),
 PQR
 As
 (
 SELECT EB.[MerchantOwnerID], EB.[FirstName], EB.[LastName],EB.[BusinessTitle],
 EB.[OwnershipPercentage], EB.[DateOfBirth],E.[EntityID],
  row_number() over(partition by E.EntityID order by E.EntityID desc) as rn
    FROM [dbo].[Merchant_Owners] EB
    JOIN [Entities] E ON E.[SourcePrimaryKey] = EB.[MerchantOwnerID] and [EntityTypeID]=2
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
 )

SELECT STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
            FROM CTE
            WHERE [EntityID] = c.[EntityID]
            ORDER BY [EmailTypeID]
            FOR XML PATH ('')),1,4,'') [EmailAddress],
   c.[EntityID],
   STUFF((SELECT '::::' + CONVERT(NVARCHAR, [PhoneType]) + '_' + [PhoneNumber]
            FROM ABC
            WHERE [EntityID] = c2.[EntityID]
            ORDER BY [PhoneType]
            FOR XML PATH ('')),1,4,'') [PhoneNumber],
   c2.[EntityID] as pid,
  c3.[City],c3.[State],c3.[EntityID], c4.[FirstName]

FROM CTE c  FULL JOIN ABC c2 ON c.EntityID = c2.EntityID,
GHK c3, PQR c4
--FULL JOIN
--GHK c3 ON c.EntityID = c3.EntityID

  WHERE c.rn = 1 OR c2.rn = 1 or c3.rn=1or c4.rn=1
  group by  c.[EntityID],c2.[EntityID],c3.EntityID,c4.EntityID

它抛出错误 Column 'GHK.City' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

如果我删除此字段，则会在GHK.State处抛出错误，依此类推......

我的CTE（GHK）已经返回7行，所以我认为他们不需要聚合它。

知道如何到处走走。

我的CTE（PQR）也返回了distict行。我也得到同样的错误。

Answer 1

根据您期望的输出，其中一项可行。

如果您期望每个EmailTypeID的个人结果（例如12 ::: 13,12 ::: 14,13 ::: 14），那么以下内容将起作用。

DECLARE @MyTableVariable TABLE([EmailAddress] NVARCHAR(100),[EmailTypeID] INT, [EntityID] INT)
INSERT @MyTableVariable VALUES ('abc@abc.com',12,101)
,('pqr@pqr.com',13,101)
,('koi@koi.com',12,102)
,('poi@poi.com',13,102)
,('blah@blah.com',14,102)
SELECT CONVERT(NVARCHAR, T1.[EmailTypeID]) + '_' + T1.[EmailAddress] + '::::' + CONVERT(NVARCHAR, T2.[EmailTypeID]) + '_' + T2.[EmailAddress] [EmailAddress]
    , T1.[EntityID]
FROM @MyTableVariable T1
JOIN @MyTableVariable T2 ON T1.[EntityID] = T2.[EntityID] AND T2.[EmailTypeID] > T1.[EmailTypeID]

如果您只想要一个连接，那么这将起作用：

DECLARE @MyTableVariable2 TABLE([EmailAddress] NVARCHAR(100),[EmailTypeID] INT, [EntityID] INT)
INSERT @MyTableVariable2 VALUES ('abc@abc.com',12,101)
,('pqr@pqr.com',13,101)
,('koi@koi.com',12,102)
,('poi@poi.com',13,102)
,('blah@blah.com',14,102)
SELECT STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
            FROM @MyTableVariable2
            WHERE [EntityID] = T1.[EntityID]
            ORDER BY [EmailTypeID]
            FOR XML PATH ('')),1,4,'') [EmailAddress]
    , [EntityID]
FROM @MyTableVariable2 T1
GROUP BY [EntityID]

作为旁注，我认为您的查询看起来更像是一系列连接而不是一系列IN语句。 e.g。

SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID]
FROM [dbo].[Email_Addresses] EA
JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
JOIN (SELECT [MechantOwnerID] FROM [dbo].[MechantOwners] WHERE [MechantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]

编辑（一起显示整个查询）：

; WITH CTE AS (
    SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID]
    FROM [dbo].[Email_Addresses] EA
    JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
    JOIN (SELECT [MechantOwnerID] FROM [dbo].[MechantOwners] WHERE [MechantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID])
SELECT STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
            FROM CTE
            WHERE [EntityID] = T1.[EntityID]
            ORDER BY [EmailTypeID]
            FOR XML PATH ('')),1,4,'') [EmailAddress]
    , [EntityID]
FROM CTE T1
GROUP BY [EntityID]

如果您更喜欢使用当前查询（或者由于某些原因我的查询不起作用），只需编辑CTE中的内容即可。

EDIT2（添加了电话号码）：

; WITH CTE AS (
    SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID]
    FROM [dbo].[Email_Addresses] EA
    JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
),
ABC
As
(
    SELECT EB.[PhoneNumber], EB.[EntityID], EB.[PhoneType]
    FROM [dbo].[Phones] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
)


SELECT STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
        FROM CTE
        WHERE [EntityID] = T1.[EntityID]
        ORDER BY [EmailTypeID]
        FOR XML PATH ('')),1,4,'') [EmailAddress],
        T1.[EntityID],
        STUFF((SELECT '::::' + CONVERT(NVARCHAR, [PhoneType]) + '_' + [PhoneNumber]
        FROM ABC
        WHERE [EntityID] = T1.[EntityID]
        ORDER BY [PhoneType]
        FOR XML PATH ('')),1,4,'') [PhoneNumber]
FROM CTE T1
LEFT JOIN ABC T2 ON T2.[EntityID] = T1.[EntityID]
GROUP BY T1.[EntityID]

EDIT3：

SELECT COALESCE(c.[EntityID], c2.[EntityID], c3.[EntityID], c4.[EntityID]) [EntityID]
    , STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
        FROM CTE
        WHERE [EntityID] = c.[EntityID]
        ORDER BY [EmailTypeID]
        FOR XML PATH ('')),1,4,'') [EmailAddress]
    , STUFF((SELECT '::::' + CONVERT(NVARCHAR, [PhoneType]) + '_' + [PhoneNumber]
        FROM ABC
        WHERE [EntityID] = c2.[EntityID]
        ORDER BY [PhoneType]
        FOR XML PATH ('')),1,4,'') [PhoneNumber]
    , c3.[City], c3.[State], c4.[FirstName]
FROM CTE c  
FULL JOIN ABC c2 ON c.EntityID = c2.EntityID
FULL JOIN (SELECT * FROM GHK WHERE RN = 1) c3 ON c3.[EntityID] = c.[EntityID] OR c3.[EntityID] = c2.[EntityID]
FULL JOIN (SELECT * FROM PQR WHERE RN = 1) c4 ON c4.[EntityID] = c.[EntityID] OR c4.[EntityID] = c2.[EntityID] OR c4.[EntityID] = c3.[EntityID]
GROUP BY c.[EntityID], c2.[EntityID], c3.[EntityID], c4.[EntityID], c3.[City], c3.[State], c4.[FirstName]

作为使用UNION如何使用当前查询的示例，这应该有效（并且避免需要GROUP BY，因为每个entityID只生成一行）。编辑4：

; WITH CTE AS (
    SELECT EA.[EmailTypeID], EA.[EmailAddress], EA.[EntityID]
    FROM [dbo].[Email_Addresses] EA
    JOIN [Entities] E ON E.[EntityID] = EA.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
    )
, ABC AS (
    SELECT EB.[PhoneNumber], EB.[EntityID], EB.[PhoneType]
    FROM [dbo].[Phones] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
    )
, GHK AS (
    SELECT EB.[Address1], EB.[City], EB.[State],EB.[EntityID],
        ROW_NUMBER() OVER(PARTITION BY E.EntityID ORDER BY E.EntityID) AS RN
    FROM [dbo].[Addresses] EB
    JOIN [Entities] E ON E.[EntityID] = EB.[EntityID]
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
    )
, PQR AS (
    SELECT EB.[FirstName], E.[EntityID],
        -- EB.[MerchantOwnerID], EB.[LastName],EB.[BusinessTitle], EB.[OwnershipPercentage], EB.[DateOfBirth],
        ROW_NUMBER() OVER(PARTITION BY E.EntityID ORDER BY E.EntityID) AS RN
    FROM [dbo].[Merchant_Owners] EB
    JOIN [Entities] E ON E.[SourcePrimaryKey] = EB.[MerchantOwnerID] and [EntityTypeID]=2
    JOIN (SELECT [MerchantOwnerID] FROM [dbo].[Merchant_Owners] WHERE [MerchantID] = 1) M ON M.[MerchantOwnerID] = E.[SourcePrimaryKey]
    JOIN (SELECT [EntityTypeID] FROM [Entity_Types] WHERE [EntityType] = 'MerchantOwner') ET ON ET.[EntityTypeID] = E.[EntityTypeID]
    )
, EntityList AS (
    SELECT [EntityID] FROM CTE
    UNION
    SELECT [EntityID] FROM ABC
    UNION
    SELECT [EntityID] FROM GHK
    UNION
    SELECT [EntityID] FROM PQR
    )
SELECT EL.[EntityID]
    , STUFF((SELECT '::::' + CONVERT(NVARCHAR, [EmailTypeID]) + '_' + [EmailAddress]
            FROM CTE
            WHERE [EntityID] = EL.[EntityID]
            ORDER BY [EmailTypeID]
            FOR XML PATH ('')),1,4,'') [EmailAddress]
    , STUFF((SELECT '::::' + CONVERT(NVARCHAR, [PhoneType]) + '_' + [PhoneNumber]
            FROM ABC
            WHERE [EntityID] = EL.[EntityID]
            ORDER BY [PhoneType]
            FOR XML PATH ('')),1,4,'') [PhoneNumber]
    , c3.[City], c3.[State], c4.[FirstName]
FROM EntityList EL
LEFT JOIN CTE c1 ON c1.[EntityID] = EL.[EntityID]
LEFT JOIN ABC c2 ON c2.[EntityID] = EL.[EntityID]
LEFT JOIN (SELECT * FROM GHK WHERE RN = 1) c3 ON c3.[EntityID] = EL.[EntityID]
LEFT JOIN (SELECT * FROM PQR WHERE RN = 1) c4 ON c4.[EntityID] = EL.[EntityID]
ORDER BY EL.[EntityID]

如何在where子句中使用动态结果连接两行或更多行？

1 个答案: