关于去重复记录的sql

现在有一个视图如图，现在我想要vessel_cn这一列不允许重复，如果重复则取seain列大的那一条记录，如果seain列的值相同则取seaout列大的那一条记录，如果seaout列值相同则取leng列大的那一条记录，不可能出现两列完全相同的情况。
比如图中第135 136行记录的vessel_cn值相同,所以应该去掉seain值小的135行。
求写sql语句对这个视图进行上述去重复选择。视图名叫v_plan

方式一：试试用max聚合：
顺序：max聚合leng-->max聚合seaout-->max聚合seain
sql如下：
select C.vessel_cn,C.seain,C.maxSeaout,C.maxLength,max(C.seain) as maxSeain from
(
select B.vessel_cn,B.seain,B.maxLength,max(B.seaout) as maxSeaout from
(
select A.vessel_cn,A.seain,A.seaout,max(A.leng) as maxLength from 你的视图 A
group by A.vessel_cn,A.seain,A.seaout
) B
group by B.vessel_cn,B.seain,B.maxLength
) C
group by C.vessel_cn,C.maxSeaout,C.maxLength

（2）方案二：根据你列出的条件用子查询过滤掉重复的记录，sql：
select A.vessel_cn,A.seain,A.seaout,A.leng from 你的视图 A
where
not exists (select 1 from 你的视图 B where B.vessel_cn=A.vessel_cn and B.seain=A.seain and B.seaout=A.seaout and B.leng>A.leng)
and
not exists (select 1 from 你的视图 C where C.vessel_cn=A.vessel_cn and C.seain=A.seain and C.seaout>A.seaout and C.leng=A.leng)
and
not exists (select 1 from 你的视图 D where D.vessel_cn=A.vessel_cn and D.seain>A.seain and D.seaout=A.seaout and D.leng=A.leng)

每行应该增加个唯一键

使用临时表试试？？
[code="java"]
create table xx_tmp as select min(cuid) as col1 from wyr_customer group by phone;
delete from wyr_customer where cuid not in (select col1 from xx_tmp);
drop table xx_tmp;
[/code]