sql – 如何从数据库表中的重复行中选择具有最新时间戳的行？

我有一张重复的桌子和三重行 – 如何选择重复但具有最新时间戳以及未加工行的行？

-------------------------------------
| pk_id | user_id | some_timestamp    |
|-------------------------------------|
| 1     | 123     | 10-Jun-12 14.30   |
| 2     | 123     | 19-Jun-12 21.50   |
| 3     | 567     | 10-Jun-12 09.23   |
| 4     | 567     | 12-Jun-12 09.45   |
| 5     | 567     | 13-Jun-12 08.40   |
| 6     | 890     | 13-Jun-12 08.44   |
 -------------------------------------

所以我最终得到：

-------------------------------------
| pk_id | user_id | some_timestamp    |
|-------------------------------------|
| 2     | 123     | 19-Jun-12 21.50   |
| 5     | 567     | 13-Jun-12 08.40   |
| 6     | 890     | 13-Jun-12 08.44   |
 -------------------------------------

解决方法

SELECT * FROM ( 
SELECT pk_id,user_id,some_timestamp,ROW_NUMBER() OVER (PARTITION BY user_id ORDER BY some_timestamp DESC) col 
FROM table) x  
WHERE x.col = 1