Need a query for grouping

Question

Hi,
I have some data and need to group it based on date. please see the below data
pkey--------from_date----------to_date------------amount-------------
1 ---------8-aug-12 ----------31-aug-12---------120
1 ---------31-aug-12 --------1-sep-12--------- 130
1 ---------1-sep-12 ----------2-sep-12--------- 150
1 ---------3-sep-12 ----------4-sep-12--------- 150
1 ---------5-sep-12 ----------7-sep-12--------- 100
1 ---------7-sep-12 ----------8-sep-12--------- 200
1 ---------8-sep-12 ----------20-sep-12---------120
1 ---------20-sep-12 --------1-oct-12--------- 130
1 ---------1-oct-12 ----------8-oct-12--------- 150
and so on.....
I need to group the data when a month finished.. e.g row 1 from_date is 08-aug-12 row 6 to_date is 08-sep-12 means almost 1 month finished so it should group and show the sum of amount with min(from_date) and max(to_Date) and total amount monhtly basis.
I am trying to make a query since many hours but no success.. the data is huge with many keys and dates ... please share your idea/tips/feedback.
Thanks
Edited by: hard_stone on Jan 18, 2013 11:38 AM

Answer 1

Try this
with t
as
select 1 pkey, to_date('08-aug-12', 'dd-mon-rr') from_date, to_date('31-aug-12', 'dd-mon-rr') to_date, 120 amount from dual
union all
select 2 pkey, to_date('31-aug-12', 'dd-mon-rr') from_date, to_date('01-sep-12', 'dd-mon-rr') to_date, 130 amount from dual
union all
select 3 pkey, to_date('01-sep-12', 'dd-mon-rr') from_date, to_date('02-sep-12', 'dd-mon-rr') to_date, 150 amount from dual
union all
select 4 pkey, to_date('03-sep-12', 'dd-mon-rr') from_date, to_date('04-sep-12', 'dd-mon-rr') to_date, 150 amount from dual
union all
select 5 pkey, to_date('05-sep-12', 'dd-mon-rr') from_date, to_date('07-sep-12', 'dd-mon-rr') to_date, 100 amount from dual
union all
select 6 pkey, to_date('07-sep-12', 'dd-mon-rr') from_date, to_date('08-sep-12', 'dd-mon-rr') to_date, 200 amount from dual
union all
select 7 pkey, to_date('08-sep-12', 'dd-mon-rr') from_date, to_date('20-sep-12', 'dd-mon-rr') to_date, 120 amount from dual
union all
select 8 pkey, to_date('20-sep-12', 'dd-mon-rr') from_date, to_date('01-oct-12', 'dd-mon-rr') to_date, 130 amount from dual
union all
select 9 pkey, to_date('01-oct-12', 'dd-mon-rr') from_date, to_date('08-oct-12', 'dd-mon-rr') to_date, 150 amount from dual
select min(from_date) from_date, max(to_date) to_date, sum(amount) amount
  from (
        select pkey, from_date, to_date, amount, next_month
          from t
         model
         dimension by (pkey)
         measures (from_date, to_date, amount, to_date('19000101', 'yyyymmdd') next_month, add_months(from_date, 1) temp_month)
         rules upsert
           next_month[any] = case when next_month[cv(pkey)-1] >= to_date[cv(pkey)] then next_month[cv(pkey)-1]
                                  else temp_month[cv(pkey)] end
group by next_month;
FROM_DATE TO_DATE   AMOUNT
08-AUG-12 08-SEP-12    850
08-SEP-12 08-OCT-12    400