Analytic function - two answers in one single select

Question

create table atest(eid varchar2(10),ukey varchar2(10),tname varchar2(40));
insert into atest values ('eid1','ukey1','tname1');
insert into atest values ('eid2','ukey1','tname2');
insert into atest values ('eid3','ukey2','tname3');
insert into atest values ('eid4','ukey2','tname3');
insert into atest values ('eid5','ukey3','tname4');
commit;I need one single sql, from which I want to find out 2 things
1. find out those ukey values for which we have multiple eids (ukey1, ukey2 in our example)
2. next, once we identify those ukeys with multiple eids, out of those, find those ukeys which have non unique tnames (in our example ukey1)
I would like to do it using analytic functions, so far tried the following
select eid, ukey,
(case when count(distinct a.eid) over (partition by (a.ukey)) >  1 then
-- do something here, above will give ukey1, ukey2, but how to identify those ukeys with non unique tnames ?
else
end) tsource
from atest a;

Answer 1

user650888 wrote:
can the same be achieved with analytics ? Yes, but why? Anyway:
select  ukey
  from  (
         select  ukey,
                 count(distinct eid) over(partition by ukey) eid_cnt,
                 count(distinct tname) over(partition by ukey) tname_cnt,
                 row_number() over(partition by ukey order by 1) rn
           from  atest
  where eid_cnt > 1
    and tname_cnt > 1
    and rn = 1
UKEY
ukey1
SQL> SY.