Question

我有下表：

ONBackup表：

Contract    FromDate    Invoice     Data
232         12/12/2017  123
232         14/02/2018  123
232         15/07/2018  123
232         14/02/2017  676
311         12/12/2017  881

“重复”的行很多，对我来说重复的是发票号相同，即其他字段可以不同。

该表有140万行（大约有100万行重复），所以不确定以下内容是否可以工作，因为我无聊地等待3个小时并开始计数，它肯定比我要占用更多CPU。< / p>

DELETE FROM ONBackup
WHERE Invoice NOT IN
(
    SELECT MIN(Invoice)
    FROM ONBackup
    GROUP BY Invoice
)

有没有一种更快的方法可以起作用？

Answer 1

使用row_number()函数：

delete b
from (select b.*, row_number() over (partition by b.invoice order by b.fromdate desc) as seq 
      from ONBackup b
     ) b
where seq > 1;

这将为每个fromdate留下最新的invoice。

Answer 2

在这里，我认为CTE是一个不错的选择：（请注意，您必须在前面的语句中使用分号结束）。

WITH CTE AS 

(
  SELECT Invoice, ROW_NUMBER() OVER (PARTITION BY INVOICE ORDER BY SELECT '1') AS RowNumb
  FROM ONBackup
)

DELETE FROM CTE WHERE RowNumb > 1

Answer 3

  DELETE A
FROM
(
  select  *,row_number() over (partition by invoice order by invoice)as rn from 
  table1
) A
WHERE A.rn > 1

删除重复项，但各保留一行

3 个答案: