mysql重复记录取最后一条记录方法

如下表：

代码如下	复制代码
CREATE TABLE `t1` ( `userid` INT(11) DEFAULT NULL, `atime` datetime DEFAULT NULL, KEY `idx_userid` (`userid`) ) ENGINE=InnoDB DEFAULT CHARSET=utf8；

数据如下：

代码如下

复制代码

MySQL> SELECT * FROM t1;

+--------+---------------------+
| userid | atime |
+--------+---------------------+
| 1 | 2013-08-12 11:05:25 |
| 2 | 2013-08-12 11:05:29 |
| 3 | 2013-08-12 11:05:32 |
| 5 | 2013-08-12 11:05:34 |
| 1 | 2013-08-12 11:05:40 |
| 2 | 2013-08-12 11:05:43 |
| 3 | 2013-08-12 11:05:48 |
| 5 | 2013-08-12 11:06:03 |
+--------+---------------------+
8 ROWS IN SET (0.00 sec)

其中userid不唯一，要求取表中每个userid对应的时间离现在最近的一条记录．初看到一个这条件一般都会想到借用临时表及添加主建借助于join操作之类的．
给一个简方法：

代码如下

复制代码

MySQL> SELECT userid,substring_index(group_concat(atime ORDER BY atime DESC),",",1) AS atime FROM t1 GROUP BY userid;

+--------+---------------------+
| userid | atime |
+--------+---------------------+
| 1 | 2013-08-12 11:05:40 |
| 2 | 2013-08-12 11:05:43 |
| 3 | 2013-08-12 11:05:48 |
| 5 | 2013-08-12 11:06:03 |
+--------+---------------------+
4 ROWS IN SET (0.03 sec)

查询及删除重复记录

删除表中多余的重复记录，重复记录是根据单个字段（peopleId）来判断，只留有rowid最小的记录

代码如下	复制代码
delete from people where peopleId in (select peopleId from people group by peopleId having count(peopleId) > 1) and rowid not in (select min(rowid) from people group by peopleId having count(peopleId )>1)

3、查找表中多余的重复记录（多个字段）

代码如下	复制代码
select * from vitae a where (a.peopleId,a.seq) in (select peopleId,seq from vitae group by peopleId,seq having count(*) > 1)

4、删除表中多余的重复记录（多个字段），只留有rowid最小的记录

代码如下	复制代码
delete from vitae a where (a.peopleId,a.seq) in (select peopleId,seq from vitae group by peopleId,seq having count() > 1) and rowid not in (select min(rowid) from vitae group by peopleId,seq having count()>1)

5、查找表中多余的重复记录（多个字段），不包含rowid最小的记录

代码如下	复制代码
select * from vitae a where (a.peopleId,a.seq) in (select peopleId,seq from vitae group by peopleId,seq having count() > 1) and rowid not in (select min(rowid) from vitae group by peopleId,seq having count()>1)

补充：数据库,mysql教程