Question

我想根据条件从表中删除重复的行。

如，表ABC：

ID  Name    Pin Status
111 Pope    909 New
222 Jazz    909 New
333 Volk    767 Assigned
444 Audi    767 New
555 Pink    435 New

我想删除具有相同＆＃39; Pin＆＃39;的重复记录。基于以下条件：

如果两个相同的Pin记录都具有新状态 - 删除其中任何一个。
如果其中一个相同的Pin记录具有“新状态”而另一个具有“已分配状态” - 请删除“新建”状态。状态记录。

我写了下面的SQL，但我想知道是否有更好的方法来处理这个使用PK。

1：

DELETE FROM ABC WHERE ID IN (
SELECT
  a.id
FROM ABC a
WHERE a.rowid <
  (SELECT max(rowid)
  FROM  ABC  b
  WHERE b.pin = a.pin
  AND a.status  = b.status
  AND b.status  in ('New')
  )
);

2：

DELETE FROM ABC WHERE ID IN (
SELECT
  a.id
FROM ABC a
WHERE a.rowid <
  (SELECT max(rowid)
  FROM  ABC  b
  WHERE b.pin = a.pin
  AND a.status  <> b.status
  AND a.status  in ('New')
  AND b.status  in ('Assigned')
  )
);

Answer 1

你可以试试这个。

--Query 1:
DELETE FROM ABC
WHERE ID NOT IN (SELECT MIN(ID)
                 FROM ABC 
                 GROUP BY Pin
                 HAVING COUNT(*)=COUNT(CASE WHEN status = 'New' THEN 1 END)
                );

--Query 2:
DELETE FROM ABC
WHERE ID IN (SELECT ID
             FROM ABC A
             WHERE status = 'New'
             AND EXISTS (SELECT 1 FROM ABC B 
                         WHERE A.Pin=B.Pin AND B.status='Assigned')

            );

Answer 2

如果每个pin最多有两条记录：

DELETE FROM ABC
WHERE rowid IN (SELECT MIN(CASE WHEN a.status = 'NEW' then rowid END) 
                FROM ABC a
                WHERE abc.pin = a.pin AND
                      a.status IN ('New', 'Assigned'
                GROUP BY a.pin
                HAVING COUNT(*) > 1
               );

Answer 3

尝试此查询 -

-- Select ID, Name, Pin, Status
Delete
From (
    Select  
        ID, Name, Pin, Status,
        Row_Number() Over(Partition By Pin Order By Pin) As PinRowNum,
        Row_Number() Over(Partition By Pin, Status Order By Status) As 
    StatusRowNum
From ABC
) As D Where StatusRowNum > 1 Or (PinRowNum > 1 And StatusRowNum = 1)

根据Oracle

3 个答案: