SQL 服务器 Pivot 没有列名

Question

我有这张桌子：

库存单位	品牌标识	条码
1	1	123
1	1	987
1	1	852
1	2	951
1	2	753
2	1	926
2	1	364
2	2	854
2	2	256
2	2	351
2	2	157
2	3	976

我需要这种查询结果格式：

库存单位	品牌标识	条码1	条码2	条码3	条码4
1	1	123	987	852
1	2	951	753
2	1	926	364
2	2	854	256	351	157
2	3	976

条形码应该是动态的，因为每个 SKU+BrandId 行都有未知的条形码

Answer 1

这个概念是建立一个动态 pivot 查询如下

 CREATE TABLE #t (sku nvarchar(5), brandid nvarchar(5), barcode nvarchar(50) )
truncate table #t
INSERT INTO #t VALUES 
    (1,1,150),
    (1,1,140),
    (1,1,111),
    (1,2,1234),
    (1,2,145),
    (2,1,345),
    (2,1,532),
    (2,2,875),
    (2,2,1237),
    (3,1,566)

select *,'Barcode'+CONVERT (nvarchar(50),ROW_NUMBER()over(partition by sku, brandid order by sku)) rn from #t   


DECLARE @colName AS NVARCHAR(MAX), @pivotQ AS NVARCHAR(max)

SELECT @colName = ISNULL(@colName + ',', '') + QUOTENAME(rn) 
FROM (select DISTINCT 'Barcode'+CONVERT (nvarchar(50),ROW_NUMBER()over(partition by sku, brandid order by sku)) rn from #t  
) AS Labels ORDER BY rn


SET @pivotQ = N'
SELECT sku,   brandid, '+@colName+'
FROM  
(
  SELECT sku,   brandid , barcode , ''Barcode''+CONVERT (nvarchar(50),ROW_NUMBER()over(partition by sku, brandid order by sku)) rn
  FROM #t
) AS SourceTable  
PIVOT  
(  
  MAX(barcode)  
  FOR rn IN ('+@colName+')  
) AS PivotTable; 
'

EXECUTE sp_executesql @pivotQ

Answer 2

我强烈建议您不要将 go 向下动态查询路由，因为它可能难以编码，并且容易受到 SQL 注入的影响。

相反，只需在正常的条件聚合中硬编码最大可能的Barcode列。

SELECT
  t.Sku,
  t.BrandId,
  Barcode1 = MAX(CASE WHEN t.rn = 1 THEN t.Barcode END),
  Barcode2 = MAX(CASE WHEN t.rn = 2 THEN t.Barcode END),
  Barcode3 = MAX(CASE WHEN t.rn = 3 THEN t.Barcode END),
  Barcode4 = MAX(CASE WHEN t.rn = 4 THEN t.Barcode END),
  Barcode5 = MAX(CASE WHEN t.rn = 5 THEN t.Barcode END),
  Barcode6 = MAX(CASE WHEN t.rn = 6 THEN t.Barcode END),
  Barcode7 = MAX(CASE WHEN t.rn = 7 THEN t.Barcode END),
  Barcode8 = MAX(CASE WHEN t.rn = 8 THEN t.Barcode END),
  Barcode9 = MAX(CASE WHEN t.rn = 9 THEN t.Barcode END),
  Barcode10 = MAX(CASE WHEN t.rn = 10 THEN t.Barcode END)
FROM  
(
    SELECT *,
      rn = ROW_NUMBER() OVER (PARTITION BY t.Sku, t.BrandId ORDER BY t.Barcode)
    FROM YourTable t
) t
GROUP BY
  t.Sku,
  t.BrandId;