将具有相同ID但不同属性的记录合并为一条记录答案

【问题标题】：Combine records with same ID, but different attributes into one record将具有相同ID但不同属性的记录合并为一条记录
【发布时间】：2017-05-05 23:53:27
【问题描述】：

例如，我有一个包含以下记录的表...

ID | Attribute
1      BLUE
1      GREEN
1      RED
2      YELLOW
2      GREEN
3      GREEN

我想把它压缩成一条包含所有属性的记录。

ID | Attribute1 | Attribute2 | Attribute3
1      BLUE         GREEN         RED
2      YELLOW       GREEN
3      GREEN

我正在走 PIVOT 的道路，但考虑到它们共享相同的 ID/Key，我不知道如何将属性明确插入到单独的列中。我正在使用 SSMS。

【问题讨论】：

您使用的是哪个DBMS？
每个ID最多可以有3个Attributes或者未知？
SQL Server 管理工作室
对于我目前正在处理的表，我拥有的最大属性是 3 个。

标签： sql sql-server tsql pivot

【解决方案1】：

试试这个

;WITH cte
     AS (SELECT *,Row_number()OVER(partition BY [ID] ORDER BY [Attribute]) rn
         FROM   Yourtable)
SELECT [ID],
       Max(CASE WHEN rn = 1 THEN [Attribute] ELSE '' END) AS [Attribute1],
       Max(CASE WHEN rn = 2 THEN [Attribute] ELSE '' END) AS [Attribute2],
       Max(CASE WHEN rn = 3 THEN [Attribute] ELSE '' END) AS [Attribute3]
FROM   cte
GROUP  BY [ID]

如果你想使用未知数量的属性，那么

DECLARE @int INT = 1,
        @cnt INT,
        @sql VARCHAR(max)

SELECT TOP 1 @cnt = Count(1)OVER(partition BY [ID])
FROM   Yourtable
ORDER  BY Count(1)OVER(partition BY [ID]) DESC

SET @sql = ';WITH cte
         AS (SELECT *,Row_number()OVER(partition BY [ID] ORDER BY [Attribute]) rn
             FROM   Yourtable)
    SELECT [ID],'

WHILE @int <= @cnt
  BEGIN
      SET @sql += 'Max(CASE WHEN rn = ' + Cast(@int AS VARCHAR(20)) + ' THEN [Attribute] ELSE '''' END) AS [Attribute' + Cast(@int AS VARCHAR(20)) + '],'
      SET @int +=1
  END

SET @sql = LEFT(@sql, Len(@sql) - 1)
SET @sql += 'FROM   cte GROUP  BY [ID]' 

exec (@sql)

【讨论】：

这很完美！我只是输入了 NULL 而不是 ' '。感谢帮助！在使用 CTE 时需要提高一点。

【解决方案2】：

如果您不需要动态，条件聚合可能会有所帮助

Select ID
      ,max(case when RN=1 then Attribute else '' end)) as Attribute1 
      ,max(case when RN=2 then Attribute else '' end)) as Attribute2 
      ,max(case when RN=3 then Attribute else '' end)) as Attribute3
 From (
        Select *
              ,RN = Row_Number() over (Partition By ID Order By Attribute)
         From YourTable
      )
 Group By ID

【讨论】：

【解决方案3】：

这应该可以工作并且使用起来非常简单

select * from(
        select 
             ID
            ,Attribute
            ,row_number() over(partition by ID order by Attribute) as AttributeNumber
        from [YourTable]
        group by ID
                ,Attribute
            )t1
PIVOT
(
MAX(Attribute)
FOR AttributeNumber in ([1],[2],[3])-- add more as needed
)piv

【讨论】：