需要SQL查询/子查询帮助

Question

我试图从SQL Server 2005上的测试信息表中获取最后10个序列号的列表。我尝试了以下操作：

SELECT DISTINCT TOP (10) Serial, DateTime
FROM [Test].[dbo].[TestInfo]
WHERE (TestedBy = 'JSMITH') ORDER BY DateTime DESC

返回重复的序列号：

+---------+-------------------------+
| Serial  | DateTime                |
+-----------------------------------+
| 1114048 | 2011-03-16 11:03:14.000 |
| 1617683 | 2011-03-11 15:07:29.000 |
| 1617683 | 2011-03-11 15:07:27.000 |
| 1617683 | 2011-03-11 15:07:26.000 |
| 1617683 | 2011-03-10 13:16:04.000 |
| 1617683 | 2011-03-10 13:15:35.000 |
| 1617683 | 2011-03-10 13:15:30.000 |
| 1617683 | 2011-03-07 13:42:48.000 |
| 1617683 | 2011-03-07 13:40:32.000 |
| 1617683 | 2011-03-07 13:37:58.000 |
+---------+-------------------------+

有没有办法使用查询或子查询来获取最后10个序列号而不重复的序列号？

Answer 1

select top (10) Serial, Max(DateTime)
from [Test].[dbo].[TestInfo]
where (TestedBy = 'JSMITH')
group by Serial
order by Max(DateTime) desc

Answer 2

SELECT TOP 10 
  * 
FROM (SELECT 
        Serial, 
        MAX(DateTime) AS DateTime
      FROM [Test].[dbo].[TestInfo]
      WHERE (TestedBy = 'JSMITH') 
      GROUP BY Serial) q1
ORDER BY q1.DateTime DESC

Answer 3

大概是这样的：

SELECT DISTINCT TOP (10) Serial, DateTime
FROM (
  SELECT Serial, MAX(DateTime) AS DateTime
  FROM [Test].[dbo].[TestInfo]
  WHERE (TestedBy = 'JSMITH') 
  GROUP BY Serial
) AS sub
ORDER BY DateTime DESC

Answer 4

它返回重复的行，因为每一行的日期时间都不相同。 如果仅需要串行字段，则必须在查询中仅写入串行字段。