Question

使用SQL Server 2008 tsql时，我试图在一组值重复时删除表中的所有记录。所以如果我的表看起来像这样：

ID COL1 COL2
1 A 1
2 A 1
3 A 2
4 乙 1
5 乙 1
6 B 2

第1,2,4,5行将全部删除。

Answer 1

;WITH d AS 
(
  SELECT col1, col2, c = COUNT(*) OVER 
    (PARTITION BY col1, col2 ORDER BY col1)
  FROM dbo.yourtable
)
DELETE d WHERE c > 1;

实际上它可以稍微更整洁：

;WITH d AS 
(
  SELECT id, c = COUNT(*) OVER 
    (PARTITION BY col1, col2 ORDER BY col1)
  FROM dbo.yourtable
)
DELETE d WHERE c > 1;

我会说，我在SQL Server 2012上测试了上述内容，但是我忘了改变SQL Server 2008的小提琴。对于SQL Server 2012之前的版本，这里有一个变体：

;WITH d AS 
(
  SELECT col1, col2
  FROM dbo.yourtable AS t
  GROUP BY col1, col2
  HAVING COUNT(*) > 1
)
DELETE t --*
FROM dbo.yourtable AS t
WHERE EXISTS 
(
  SELECT 1 FROM d 
  WHERE col1 = t.col1 AND col2 = t.col2
);

你认为只需DELETE d;就可以了，但你得到：

Msg 4403，Level 16，State 1，Line 2
无法更新视图或函数'd'，因为它包含聚合，DISTINCT或GROUP BY子句，或PIVOT或UNPIVOT运算符。

Answer 2

试试这个：

DELETE t
FROM dbo.yourTabe t
JOIN (
  SELECT col1,col2,COUNT(1) cnt
  FROM dbo.YourTable
  GROUP BY col1, col2
  HAVING COUNT(1)>1
) s
ON t.col1 = s.col1
AND t.col2 = s.col2

如何删除所有重复记录（包括原始记录）

2 个答案: