Dont want to use group by with sum function..

Question

Hi all
I am using sum function but i dont want to use group by since my query give me following output which i dont want I want.
0------------22---------(null)     Furniture
0-----------     3700------     (null)     अनपेड बिल्स देणे
15800-----14202-----(null)     Petty Cash
(null)------     (null)------ 9109     Furniture
(null)------ (null)------ 1789     Petty Cash
(null)-------(null)------ 0     Checking In-Transfer
I want like
0------------22---------9109     Furniture
0-----------     3700------     0     अनपेड बिल्स देणे
15800-----14202-----1789     Petty Cash
what should be the problem how I can handle the query my query is
select  null as totalcr, null as totaldr, (sum(acct.amtsourcecr)-sum(acct.amtsourcedr)) as opening , ev.name as accountname
from fact_acct acct right OUTER join  c_elementvalue ev on acct.account_id= ev.c_elementvalue_id
where acct.dateacct < '01/04/2010'
and ev.cr_account_base=2
group by ev.name
UNION
select sum(acct.amtsourcecr) as totalcr, sum(acct.amtsourcedr) as totaldr, null as opening
*,ev.name as accountname*
from fact_acct acct right OUTER join  c_elementvalue ev on acct.account_id= ev.c_elementvalue_id
where  acct.datetrx BETWEEN '01/04/2010' and '01/04/2010'
group by  ev.name
if I remove group by it shows me error .....not a single group by function ...
please help me out ,,,
thanking you

Answer 1

This is one way you can do it without a group by:
SQL> with t as
  2    (
  3      select 0 amt1, 22 amt2, null amt3, 'Furniture' tag from dual union all
  4      select 15800,14202,null, 'Petty Cash' from dual union all
  5      select null, null, 9109, 'Furniture' from dual union all
  6      select null, null, 1789, 'Petty Cash' from dual
  7    )
  8  select tag, amt1, amt2, amt3
  9  from
10  (
11    select sum(amt1) over (partition by tag) amt1
12          ,sum(amt2) over (partition by tag) amt2
13          ,sum(amt3) over (partition by tag) amt3
14          ,first_value(t.tag) over (partition by t.tag order by rownum) tag
15          ,row_number() over (partition by t.tag order by rownum) rn
16    from t
17  )
18  where rn = 1
19  /
TAG              AMT1       AMT2       AMT3
Furniture           0         22       9109
Petty Cash      15800      14202       1789