SQL : 将多个左连接行输出为 csv 文件的一条记录

Question

我有 3 个 tables :

Stock - ID, Description, Quantity etc
Equipment - ID, Description
LinkStockEquipment - StockID, EquipmentID, Quantity

每个库存项目可能需要多种类型的设备。

我想将所有数据输出到一个csv文件中给客户。

理想情况下他们想要：
StockID、StockDescription、StockQuantity、Equipment1、Equipment1Qty、Equipment2、Equipment2Qty、Equipment3、Equipment3Qty 等

所以我的出发点是：

SELECT * FROM Stock 
LEFT JOIN LinkStockEquipment ON LinkStockEquipment.StockID = Stock.ID 
LEFT JOIN Equipment ON Equipment.ID = LinkStockEquipment.EquipmentID

当然，当每个库存项目有多种类型的设备时，这只会给我多行。我可以使用 SORT 和 MAX 或 MIN 来获取顶部或底部的行，但是我将如何访问剩余的行，一个接一个地 link 它们。

会不会像（语法错误但试图传达信息）

SELECT * FROM Stock 
LEFT JOIN LinkStockEquipment AS LSE1 ON LSE1.StockID = Stock.ID LIMIT 1
LEFT JOIN LinkStockEquipment AS LSE2 ON LSE2.StockID = Stock.ID AND LSE1.StockID <> LSE2.StockID LIMIT 1
LEFT JOIN LinkStockEquipment AS LSE3 ON LSE3.StockID = Stock.ID AND LSE1.StockID <> LSE2.StockID AND LSE2.StockID <> LSE3.StockID LIMIT 1
LEFT JOIN Equipment ON Equipment.ID = LinkStockEquipment.EquipmentID

很抱歉，如果这已经被直接回答了，我之前只使用过基本的SQL，所以我没有完全理解类似问题的其他答案。

MySQL Left Join Many to One Row可能会做我想做的事，但我还需要每种设备的数量，所以这似乎不是解决方案。

更新 1： 我也查看了 MySQL pivot row into dynamic number of columns，但是当我每次都需要设备和数量而不是只需要一个项目时，我没有看到它如何应用。我会阅读更多关于 Pivoting 的内容。

更新二： 阅读有关 pivoting 的信息，它会为每种设备类型提供一个单独的列，但可能有 100 种不同的设备类型。但是，在这 100 种不同类型中，每个库存商品可能只有 2 种或 3 种。因此我只想列出那些设备类型，而不是每个列。

给定：

Stock ID | Equipment ID | Equipment Qty
---------------------------------------
A1       | E1           | 5
A1       | E2           | 3
A2       | E3           | 4
A2       | E4           | 6
A3       | E5           | 2

我愿意

Stock ID | Eqpmnt ID 1 | Eqpmnt Qty 1 | Eqpmnt ID 2 | Eqpmnt Qty 2
------------------------------------------------------------------
A1       | E1          | 5            | E2          | 3
A2       | E3          | 4            | E4          | 6
A3       | E5          | 2            |             |

没有

Stock ID | E1 | E2 | E3 | E4 | E5 
------------------------------------------------------------------
A1       | 5  | 3  |    |    |
A2       |    |    | 4  | 6  |
A3       |    |    |    |    | 2

除非我误解了旋转？

更新和结果： 出于某种原因，我无法将其设置为问题的答案，但是万一这对其他人有帮助...

感谢下面的 Juan，我使用这个问题的信息在我的 LinkStockEquipment table 中添加了一行：

UPDATE t1
SET t1.Counter = t2.Counter
FROM  LinkStockEquipment t1
LEFT JOIN
(
    SELECT
    t.EquipmentID,
    t.StockID ,
    (SELECT COUNT(*) FROM [LinkStockEquipment] AS x WHERE x.EquipmentID <= t.EquipmentID
    AND x.StockID = t.StockID) AS Counter
    FROM [LinkStockEquipment] t
) as t2
    ON t1.EquipmentID= t2.EquipmentID AND t1.StockID = t2.StockID

那么 table 看起来像：

Stock ID | Equipment ID | Equipment Qty | Counter
-------------------------------------------------
A1       | E1           | 5             | 1
A1       | E2           | 3             | 2
A2       | E3           | 4             | 1
A2       | E4           | 6             | 2
A3       | E5           | 2             | 1

现在我可以旋转了，所以使用

SELECT P.StockID, P.[1], L1.Qty, P.[2], L2.Qty, P.[3], L3.Qty
FROM (SELECT StockID, EquipmentID, Counter
FROM [LinkStockEquipment]) AS D
PIVOT(MAX(EquipmentID) FOR Counter IN([1],[2],[3])) AS P
LEFT JOIN [LinkStockEquipment] L1 ON L1.StockID = P.StockID
AND L1.EquipmentID = P.[1]
LEFT JOIN [LinkStockEquipment] L2 ON L2.StockID = P.StockID 
AND L2.EquipmentID = P.[2]
LEFT JOIN [LinkStockEquipment] L3 ON L3.StockID = P.StockID 
AND L3.EquipmentID = P.[3]

我可以加入 Stock table 详细信息，我得到了想要的结果。

Answer 1

如果您有固定数量的库存设备（我假设为 3 个）

您可以使用 GROUP BY 进行转轴。

SQL DEMO

WITH tmpResult as (
    SELECT 
         [Stock ID],
         'Equipment ID ' + CAST(rn AS VARCHAR(16)) as lblEq,
         [Equipment ID],
         [Equipment Qty]
    FROM ( SELECT *, row_number() over (partition by [Stock ID] ORDER BY [Equipment ID]) as rn
           FROM LinkStockEquipment ) as rows
)    
SELECT [Stock ID], 
       MAX( CASE WHEN lblEq = 'Equipment ID 1' THEN [Equipment ID]  END) as [Eqp ID 1],
       MAX( CASE WHEN lblEq = 'Equipment ID 1' THEN [Equipment Qty] END) as [Eqp Qty 1],
       MAX( CASE WHEN lblEq = 'Equipment ID 2' THEN [Equipment ID]  END) as [Eqp ID 2],
       MAX( CASE WHEN lblEq = 'Equipment ID 2' THEN [Equipment Qty] END) as [Eqp Qty 2],
       MAX( CASE WHEN lblEq = 'Equipment ID 3' THEN [Equipment ID]  END) as [Eqp ID 3],
       MAX( CASE WHEN lblEq = 'Equipment ID 3' THEN [Equipment Qty] END) as [Eqp Qty 3]
FROM       tmpResult
GROUP BY [Stock ID];

输出

现在，如果您想使用 PIVOT，重要的部分是数据准备。在这种情况下，我必须将数量转换为字符串。同样，您需要知道要转换的字段数

SQL DEMO

WITH tmpResult as (
    SELECT 
         [Stock ID],
         'Eqp ID ' + CAST(rn AS VARCHAR(16)) as label,
         [Equipment ID] as [Value]
    FROM ( SELECT *, row_number() over (partition by [Stock ID] ORDER BY [Equipment ID]) as rn
           FROM LinkStockEquipment ) as rows

    UNION ALL

    SELECT 
         [Stock ID],
         'Eqp Qty ' + CAST(rn AS VARCHAR(16)) as label,
         CAST([Equipment Qty] AS VARCHAR(16))  as [Value]
    FROM ( SELECT *, row_number() over (partition by [Stock ID] ORDER BY [Equipment ID]) as rn
           FROM LinkStockEquipment ) as rows

)    
SELECT [Stock ID], 
       [Eqp ID 1], [Eqp Qty 1],
       [Eqp ID 2], [Eqp Qty 2]       
FROM   ( SELECT * FROM tmpResult ) as x
PIVOT ( 
        max( [Value] ) 
        for label in  ( [Eqp ID 1], [Eqp Qty 1], [Eqp ID 2], [Eqp Qty 2] )
      ) as pvt

输出

现在如果你不知道你有多少设备，那么你需要动态 PIVOT。

SQL DEMO

首先你需要一个时间table。

SELECT 
     [Stock ID],
     [label],
     [Value]
INTO tmpResult
FROM (
    SELECT 
         [Stock ID],
         'Eqp ID ' + CAST(rn AS VARCHAR(16)) as label,
         [Equipment ID] as [Value]
    FROM ( SELECT *, row_number() over (partition by [Stock ID] ORDER BY [Equipment ID]) as rn
           FROM LinkStockEquipment ) as rows

    UNION ALL

    SELECT 
         [Stock ID],
         'Eqp Qty ' + CAST(rn AS VARCHAR(16)) as label,
         CAST([Equipment Qty] AS VARCHAR(16))  as [Value]
    FROM ( SELECT *, row_number() over (partition by [Stock ID] ORDER BY [Equipment ID]) as rn
           FROM LinkStockEquipment ) as rows
    ) as x;

然后你需要准备数据透视查询：

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX);

SET @cols = STUFF((SELECT distinct ',' + QUOTENAME(c.label)
            FROM tmpResult c
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'');

SELECT @cols; 

set @query = 'SELECT [Stock ID], ' + @cols + ' FROM  
            (
                SELECT *
                FROM tmpResult
           ) x
            pivot 
            (
                 max(Value)
                for label in (' + @cols + ')
            ) p '

execute(@query);

输出

这里的问题是列顺序。我试试看能不能修好。

SQL : 将多个左连接行输出为 csv 文件的一条记录

SQL : Output multiple left joined rows as one record for csv file

sql-server

join

sql-limit