sql-server - 从表中获取以逗号分隔的一组值，其中另一个表上的另一个引用值出现两次(或更多)

Question

假设 SQL Server 2014 中的数据库设置如下:

DECLARE @MATERIAL TABLE (ID int, CODE varchar(30));

INSERT @MATERIAL (ID, CODE) VALUES
(1, 'D3033MBBY'),
(2, 'D3033MBTY'),
(3, '011130-01'),
(4, '011130-04C'),
(5, '021002'),
(6, '021017-B'),
(7, '021134-01'),
(8, '021135-01'),
(9, '021955-01'),
(10, '3LS91101-550'),
(11, 'D3049MBRB'),
(12, 'EF0118'),
(13, 'FV8130'),
(14, 'FY7009'),
(15, 'H05802'),
(16, 'D3033MRTE');

DECLARE @SUBSTITUTE TABLE (ID int, ITEID int, SUBSTITUTECODE varchar(100));

INSERT @SUBSTITUTE (ID, ITEID, SUBSTITUTECODE) VALUES
(5232, 1, '191045762418'),
(5442, 2, '191045762418'),
(6435, 3, '5206432380030'),
(6573, 4, '5206432380030'),
(6582, 5, '5206432357131'),
(6683, 6, '5206432369486'),
(7332, 7, '5206432380610'),
(7482, 8, '5206432380818'),
(7721, 9, '5206432346029'),
(7831, 10, '5205172116350'),
(8034, 11, '191045480992'),
(8184, 12, '4061622759543'),
(8284, 13, '4062058577497'),
(8573, 14, '4064039588089'),
(9438, 15, '4064048672519'),
(9746, 16, '191045762418');

SELECT sub.SUBSTITUTECODE
FROM @SUBSTITUTE AS sub
INNER JOIN @MATERIAL AS prod ON prod.ID = sub.ITEID
GROUP BY sub.SUBSTITUTECODE
HAVING COUNT(sub.SUBSTITUTECODE) > 1;

我想创建一个会产生以下结果集的查询:

<表类="s-表"><头>代码替换代码<正文>D3033MBBY,D3033MBTY,D3033MRTE191045762418011130-01,011130-04C5206432380030

换句话说，我想要一组以逗号分隔的 CODE在@MATERIAL哪里有重复的 SUBSTITUTECODE表 @SUBSTITUTE 中这些记录的引用

间接地，我可以找到 CODE与那些重复的 SUBSTITUTECODE 相对应的 s带有以下查询:

SELECT prod.CODE, sub.SUBSTITUTECODE
FROM @SUBSTITUTE AS sub
INNER JOIN @MATERIAL AS prod ON prod.ID = sub.ITEID
WHERE sub.SUBSTITUTECODE IN (SELECT sub.SUBSTITUTECODE
    FROM @SUBSTITUTE AS sub
    INNER JOIN @MATERIAL AS prod ON prod.ID = sub.ITEID
    GROUP BY sub.SUBSTITUTECODE
    HAVING COUNT(sub.SUBSTITUTECODE) > 1)

可以找到上述案例的工作 fiddle here .

请注意，此场景的完整案例在 SQL Server 2014 上运行。

TIA

Answer 1

不错的入门 fiddle ，谢谢!如果我们只是把你已经拥有的东西放在 CTE 中，我们可以围绕它编写一个标准的字符串聚合:

;WITH subs AS 
(
  SELECT prod.CODE, sub.SUBSTITUTECODE
  FROM @SUBSTITUTE AS sub
  INNER JOIN @MATERIAL AS prod ON prod.ID = sub.ITEID
  WHERE sub.SUBSTITUTECODE IN (SELECT sub.SUBSTITUTECODE
    FROM @SUBSTITUTE AS sub
    INNER JOIN @MATERIAL AS prod ON prod.ID = sub.ITEID
    GROUP BY sub.SUBSTITUTECODE
    HAVING COUNT(sub.SUBSTITUTECODE) > 1)
)
SELECT CODES = STUFF((SELECT ',' + CODE 
  FROM subs AS s2 WHERE s2.SUBSTITUTECODE = subs.SUBSTITUTECODE
  FOR XML PATH(''), TYPE).value(N'./text()[1]', N'nvarchar(max)'),1,1,''),
    SUBSTITUTECODE FROM subs
  GROUP BY SUBSTITUTECODE;

• 示例 db<>fiddle

但我们可以稍微简化这段代码，最重要的是避免引用两个表两次，如下所示:

;WITH subs AS
(
  SELECT s.ITEID, s.SUBSTITUTECODE, m.CODE, 
    c = COUNT(*) OVER (PARTITION BY s.SUBSTITUTECODE)
  FROM @SUBSTITUTE AS s
  INNER JOIN @MATERIAL AS m
  ON m.ID = s.ITEID
)
SELECT CODES = STUFF((SELECT ',' + CODE
  FROM subs AS s2 WHERE s2.SUBSTITUTECODE = subs.SUBSTITUTECODE
  FOR XML PATH(''), TYPE).value(N'./text()[1]', N'nvarchar(max)'),1,1,''),
    SUBSTITUTECODE
  FROM subs 
  WHERE c > 1 
  GROUP BY SUBSTITUTECODE;

• 示例 db<>fiddle

请注意，在更现代的 SQL Server (2017+) 版本上，STRING_AGG() 使这更容易:

SELECT CODES = STRING_AGG(m.CODE, ','), s.SUBSTITUTECODE
  FROM @SUBSTITUTE AS s
  INNER JOIN @MATERIAL AS m
  ON m.ID = s.ITEID
  GROUP BY s.SUBSTITUTECODE
  HAVING COUNT(*) > 1;

• 示例 db<>fiddle