SQL Server查詢以查找具有聚合功能的記錄在一列中但在select子句中有多列

Question

這是Customer表的最小化版本。 可能有具有相同帳號的客戶映射到不同的Group。 我正在尋找找出映射到多個組的客戶編號。 當我使用sybase時，下面的查詢工作正常。 相同查詢在SQL Server中不起作用。

我可以在以下一個查詢中同時獲取custAccnt和相應的custId。

select DISTINCT lt.custAccnt, lt.custId from VAL_CUSTOMERS lt 
where lt.eligible = 'Y' group by lt.custAccnt 
having count(distinct lt.custId) > 1

+----------+-----------+---------+----------+
| custName | custAccnt | custId  | eligible |
+----------+-----------+---------+----------+
| Joe      | AB1VU1235 |   43553 | Y        |
| Joe      | AB1VU1235 |  525577 | Y        |
| Lucy     | CDNMY4568 |  332875 | Y        |
| Lucy     | CDNMY4568 |  211574 | Y        |
| Lucy     | CDNMY4568 |  211345 | Y        |
| Manie    | TZMM7S009 |  123890 | Y        |
| Tom      | YFDU1235  | 1928347 | Y        |
| Tom      | YFDU1235  |  204183 | Y        |
| Chef     | TNOTE6573 |  734265 | Y        |
+----------+-----------+---------+----------+

結果：-

+-----------+---------+
| AB1VU1235 |   43553 |
| AB1VU1235 |  525577 |
| CDNMY4568 |  332875 |
| CDNMY4568 |  211574 |
| CDNMY4568 |  211345 |
| YFDU1235  | 1928347 |
| YFDU1235  |  204183 |
+-----------+---------+

Answer 1

我認為這可能有效...

“ ...映射到多個組的客戶編號...”，<- group是custAcct嗎？

select t.custAccnt, t.custId 
from VAL_CUSTOMERS t 
where (Select count(distinct custAccnt ) 
       from VAL_CUSTOMERS 
       Where custId = t.custId) > 1

語句“ ...映射到多個組的客戶編號...”沒有提及“資格”，因此我沒有提及。 如果您真的要說：

“ ...映射到多個組的合格客戶編號...”，然后嘗試以下操作：

select t.custAccnt, t.custId 
from VAL_CUSTOMERS t 
where eligible = 'Y'
  and (Select count(distinct custAccnt ) 
       from VAL_CUSTOMERS 
       Where custId = t.custId) > 1

或者，這可能會更快...它回答的問題略有不同，但是（我認為）同等的問題，
“找到…… 合格的客戶編號，其中同一客戶編號的另一行已映射到不同的custAccnt ...”

select t.custAccnt, t.custId 
from VAL_CUSTOMERS t 
where eligible = 'Y'
  and exists 
      (Select * from VAL_CUSTOMERS 
       Where custId = t.custId
         and custAccnt != t.custAccnt ) 

Answer 2

有很多方法可以解決這個問題。 這里有一些應該起作用的。

select lt.custAccnt
    , lt.custId 
from VAL_CUSTOMERS lt 
cross apply
(
    select c.custAccnt
    from VAL_CUSTOMERS c
    where c.custAccnt = lt.custAccnt
    group by c.custAccnt
    having count(*) > 1
) x
where lt.eligible = 'Y'


select lt.custAccnt
    , lt.custId 
from VAL_CUSTOMERS lt 
where lt.eligible = 'Y'
AND lt.custAccnt IN
(
    select c.custAccnt
    from VAL_CUSTOMERS c
    group by c.custAccnt
    having count(*) > 1
) 

Answer 3

如果表中的custAccnt和custId重復，則@Sean查詢將不起作用。

WITH cte AS(SELECT *
                 , COUNT (custId) OVER (PARTITION BY custAccnt) AS CntcustId
                 , ROW_NUMBER () OVER (PARTITION BY custAccnt, custId ORDER BY custName)    AS Rownum
                 FROM VAL_CUSTOMERS
                 WHERE eligible = 'Y'
)
SELECT custAccnt, custId
FROM cte
WHERE CntcustId>1
AND Rownum = 1;

使用行號消除重復項。

Answer 4

;WITH cte1
( custName , custAccnt , custId  , eligible )
As
(

 SELECT 'Joe'      ,'AB1VU1235' ,   43553 , 'Y'    UNION ALL
 SELECT 'Joe'      ,'AB1VU1235' ,  525577 , 'Y'    UNION ALL
 SELECT 'Lucy'     ,'CDNMY4568' ,  332875 , 'Y'    UNION ALL
 SELECT 'Lucy'     ,'CDNMY4568' ,  211574 , 'Y'    UNION ALL
 SELECT 'Lucy'     , 'CDNMY4568' ,  211345 , 'Y'   UNION ALL
 SELECT 'Manie'    ,'TZMM7S009' ,  123890 , 'Y'    UNION ALL
 SELECT 'Tom'      ,'YFDU1235'  , 1928347 , 'Y'    UNION ALL
 SELECT 'Tom'      ,'YFDU1235'  ,  204183 , 'Y'    UNION ALL
 SELECT 'Chef'     ,'TNOTE6573' ,  734265 , 'Y'    
  )

,cte2 AS (
    SELECT custName
          ,custAccnt
          ,count(custName) cnt
    FROM cte1
    GROUP BY custName,custAccnt
    )
,cte3 AS (
    SELECT custName
        ,cnt
    FROM cte2 WHERE cnt <> 1
         )

SELECT custAccnt
    ,custId
FROM cte1
WHERE custName IN (
        SELECT custName
        FROM cte3
        )