sql - 根据连接中的值将连接结果添加为新列

Question

我从来没有做过这样的事情，我试图在谷歌上找到它但没有结果。

我有 3 个如下所示的表：

订单：

OdredID (int) PK,
UserID (int) FK,
OdredDate (datetime)

成分：

ComponentID (int) PK,
Name (nvarchar(50)),
Type (nvarchar(max))

订单组件：

OrderComponentID (int) PK,
OrderID (int) FK,
ComponentID (int) FK,
Value (nvarchar(max))

假设一个订单有 3 个具有名称的组件：[CPU, Motherboard, Memory]和值[1GHz, AsusP5, 2GB Kingston DDR3]

我需要一个结果有这样的列：

OrderID  UserID   Date          CPU    Motherboard   Memory
   1        1     2012-05-21    1GHz   AsusP5        2GB Kingston DDR3

基本上我需要每个连接行作为新列，名称取自Name连接表的列，值取自Value列。

Answer 1

尝试这个：

SELECT
  o.orderid,
  o.userid,
  MAX(CASE WHEN c.Name = 'CPU' THEN oc.Value END) AS 'CPU',
  MAX(CASE WHEN c.Name = 'Motherboard' THEN oc.Value END) AS 'Motherboard',
  MAX(CASE WHEN c.Name = 'Memory' THEN oc.Value END) AS 'Memory'
FROM orders o
INNER JOIN ordercomponents oc ON c.orderid = oc.orderId
INNER JOIN Components c ON oc.componentid = c.componentid
GROUP BY o.orderid, o.userid

SQL 小提琴演示

请注意：这是执行此操作的标准 SQL 方式。但是您可以使用 SQL ServerPIVOT表运算符来执行相同的操作，如下所示：

SELECT  *
FROM
(
   SELECT o.orderid, o.userid, c.Name 'Name', oc.value 'value'
   FROM orders o
   INNER JOIN ordercomponent oc ON o.orderid = oc.orderId
   INNER JOIN Components c ON oc.componentid = c.componentid
 ) t 
PIVOT
(
  MAX(value)
  FOR Name IN ([CPU], [Motherboard], [Memory])
) p;

但这适用于一组预定义的值，例如[CPU], [Motherboard], [Memory].

对于未知数量的值，您必须像这样动态地执行此操作：

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

select @cols = STUFF((SELECT distinct ',' + QUOTENAME(c.name) 
                    from Components c
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query = 'SELECT orderid, userid  ' + @cols + ' from 
             (
                 select o.orderid, o.userid, c.Name Name, oc.value value
                 FROM orders o
                 INNER JOIN ordercomponent oc ON o.orderid = oc.orderId
                 INNER JOIN Components c ON oc.componentid = c.componentid
            ) x
            pivot 
            (
                max(value)
                for name in (' + @cols + ')
            ) p '

execute(@query);

更新的 SQL Fiddle 演示