sql语句：
[sql]
create table dispatch_result (
dr_id number not null,
sheet_id number,
check_next varchar(20),
check_time date,
check_sn char(8),
check_comment varchar(255),
check_status number,
constraint PK_RESULT primary key (dr_id)
); www.2cto.com
表记录：
DR_ID SHEET_ID CHECK_NEXT CHECK_TIME CHECK_SN CHECK_COMMENT CHECK_STATUS
---------- ---------- -------------------- ----------- -------- -------------------------------------------------------------------------------- ------------
1 1 10001 2012/8/12 11:11:23 10000 1
2 1 10002 2012/9/15 11:15:24 10001 2
3 2 10001 2012/8/15 9:16:10 10000 1
4 2 10002 2012/9/18 11:16:23 10001 2
查询以SHEET_ID 分组取最大时间记录：
第一种子查询方式：
[sql]
select * from dispatch_result dr where dr.check_time=(
select max(check_time) from dispatch_result where sheet_id=dr.sheet_id)
DR_ID SHEET_ID CHECK_NEXT CHECK_TIME CHECK_SN CHECK_COMMENT CHECK_STATUS
---------- ---------- -------------------- ----------- -------- -------------------------------------------------------------------------------- ------------
2 1 10002 2012/9/15 11:15:24 10001 2
4 2 10002 2012/9/18 11:16:23 10001 2
www.2cto.com
第二种group by方式:
[sql]
select t1.* from dispatch_result t1 right join
(select sheet_id,max(check_time) ct from dispatch_result group by sheet_id) t2
on t2.sheet_id=t1.sheet_id and t2.ct=t1.check_time
DR_ID SHEET_ID CHECK_NEXT CHECK_TIME CHECK_SN CHECK_COMMENT CHECK_STATUS
---------- ---------- -------------------- ----------- -------- -------------------------------------------------------------------------------- ------------
2 1 10002 2012/9/15 11:15:24 10001 2
4 2 10002 2012/9/18 11:16:23 10001 2
或
[sql]
select * from dispatch_result
inner join (select max(dr.check_time) as check_time, dr.sheet_id from dispatch_result dr group by dr.sheet_id) t1
using(check_time,sheet_id);
www.2cto.com
CHECK_TIME SHEET_ID DR_ID CHECK_NEXT CHECK_SN CHECK_COMMENT CHECK_STATUS
----------- ---------- ---------- -------------------- -------- -------------------------------------------------------------------------------- ------------
2012/9/15 11:15:24 1 2 10002 10001 2
2012/9/18 11:16:23 2 4 10002 10001 2
第三种partition by方式:
[sql]
select t2.* from
(select t1.*,row_number() over
(partition by t1.sheet_id order by t1.check_time desc nulls last) rn from dispatch_result t1) t2
where rn=1
www.2cto.com
DR_ID SHEET_ID CHECK_NEXT CHECK_TIME CHECK_SN CHECK_COMMENT CHECK_STATUS
---------- ---------- -------------------- ----------- -------- -------------------------------------------------------------------------------- ------------
2 1 10002 2012/9/15 11:15:24 10001 2
4 2 10002 2012/9/18 11:16:23 10001 2