How to flatten rows

Question

Hi All,
I am trying to build a fact table.
The fee snapshot table is as follows
id,
date,
expense_type_cd,
cost_type_cd,
cost_life_cycle_cd,
fee_amt
My fact tables has columns like amt1,amt2,amt3,amt4,amt5
amt1= fee_amt from fee snapshot table where expense_type_cd in (x,y) and cost_type_cd in (q)
amt2= fee_amt from fee snapshot table where expesene_type_cd in (y)
amt3= fee_amt from fee snapshot table where cost_life_cycle_cd in (z)
create table temp as
select id, expense_ty_cd,cost_ty_cd, cost_life_cycle_cd, sum(fee_amt) amt
from fee_snpsht
group by id,rollup(expense_ty_id,cost_ty_cd,cost_life_cycle_cd)
the amounts are based on different combinations of there is no fixed combination
I created a temp table and calculated the amts using cube and rollup functions..but my question here is, to populate each and every column, do i need to write a subquery on the snapshot. i.e
insert into fact_table
(id,amt1,amt2,amt3,)
select
a.id,
(select fee_amt from temp where expense_ty_cd in(x,y) and cost_type_cd in (q) and temp.id=a.id) amt1,
(select fee_amt from temp where expense_ty_cd in(y) and cost_life_cycle_cd in (q,x,y) and temp.id=a.id) amt2,
(select fee_amt from temp where cost_type_cd in (q) and temp.id=a.id) amt3,
from fee_snapshot a
I have 20 such amount fileds. So for every amt field do I need to access the table in the above way or is there an efficient way to do this?
Please let me know i did not make sense.
your help is greatly appreaciated.
Thanks
Haro

Answer 1

You can do it directly. NOT TESTED !
Just to convey the idea clearly
select id,date,sum(fee_amt) fee_amt
  from fee_snpsht
where instr(:expense_type_cd,expense_type_cd) > 0
   and instr(:cost_type_cd,cost_type_cd) > 0
   and instr(:cost_life_cycle_cd,cost_life_cycle_cd) > 0
group by idif the expense_type_cd,cost_type_cd and cost_life_cycle_cd values are not varchar2(1) you must change all instr to (comma as list separator assumed)
instr(',' || :cost_type_cd || ',',',' || cost_type_cd || ',') > 0you must also set up a what does stand for any_value (null in this example)
instr(',' || nvl(:cost_type_cd,cost_type_cd) || ',',',' || cost_type_cd || ',') > 0something to play with:
select id,date,sum(fee_amt) fee_amt
  from fee_snpsht
where instr(',' || nvl(:expense_type_cd,expense_type_cd) || ',',',' || expense_type_cd || ',') > 0
   and instr(',' || nvl(:cost_type_cd,cost_type_cd) || ',',',' || cost_type_cd || ',') > 0
   and instr(',' || nvl(:cost_life_cycle_cd,cost_life_cycle_cd) || ',',',' || cost_life_cycle_cd || ',') > 0
group by idRegards
Etbin