这个 Sql可以用 DISTINCT实现吗

[likesome]

表1 字段n1,n2 n3,记录如下

a1 k1 100
a1 k1 200
a1 k2 100
b1 k1 100
b1 k1 200
b1 k2 200

sql n1,n2 sum(n3) as n2 from 表1 group by n1,n2 order by n1,n2
结果如下

a1 k1 300
a1 k2 100
b1 k1 300
b1 k1 200

我现要求去掉a1的重复行,即结果如下

a1 k1 300
k2 100
b1 k1 300
k1 200

如果sql不能实现,在activereport中能实现码,怎样实现

[likesome]

上面有点写错

表1 字段n1,n2 n3,记录如下

a1 k1 100
a1 k1 200
a1 k2 100
b1 k1 100
b1 k1 200
b1 k2 200

sql n1,n2 sum(n3) as n2 from 表1 group by n1,n2 order by n1,n2
结果如下

a1 k1 300
a1 k2 100
b1 k1 300
b1 k2 200

我现要求去掉n1的重复行,即结果如下

a1 k1 300
k2 100
b1 k1 300
k2 200

如果sql不能实现,在activereport中能实现码,怎样实现

[friendliu]

select n1,n2,min(n3) from tb group by n1,n2

[pisces007]

create table test(n1 varchar(20),n2 varchar(20),n3 int)
go
insert test
select 'a1','k1',100
union all select 'a1','k1',200
union all select 'a1','k2',100
union all select 'b1','k1',100
union all select 'b1','k1',200
union all select 'b1','k2',200

select identity(int,1,1) as id,n1,n2 ,sum(n3) as n3
into #a from test group by n1,n2 order by n1,n2
select (case when L.id=R.id then L.n1 else '' end) as n1,L.n2,L.n3
from #a L left join (select min(id) as id from #a group by n1) R on L.id=R.id

drop table #a
drop table test

/*
测试结果
a1 k1 300
k2 100
b1 k1 300
k2 200
*/