Anlaytical Func

Question

hi
select 1 slno, 1000 Amount , 0.12 Int_Rate, 1000*0.12 IntAmt from dual
union
select 2 slno, 1000 Amount , 0.13 Int_Rate ,1120 * 0.13 from dual
SLNO     AMOUNT     INT_RATE     INTAMT
1     1000     0.12     120
2     1000     0.13     145.6
Column 120 is the product of amount*int_rate(1000*0.12)
Column 145.6 is the product of amount*int_rate(1000*0.12)*int_rate(1000*0.13)
Can i get the help from somebody to compute using an analytical function to get the result
Thanks in Advance

Answer 1

user639304 wrote:
Hi,
That's great Sven W. In SQL, we have something to sum the values of rows, but we have nothing for the product of values of rows (or, at least, if something exist to express directly the product of numbers, we don't often use it) . So I was thinking to how use the existing functionalities of SQL in order to have the product of values in a given column, and I was thinking to the ln and exp functions. But I didn't arrive to the point where Sven W. has arrived. Sven W. has used the formula:
x1 * x2 * ... * xn = exp(ln(x1) + ln (x2) + ... + ln(xn)), in other terms, roughly speaking,
Product of values = exp(Sum of the ln of these values).Thanks for the explaination. I didn't had a proper idea how to write one myself. ;)
A minor issue with this solution is that it might be a little CPU intensive. For example I had to add rounding, because the floating point math involved led to rounding issues.
So if we need this often, it could be better to build our own aggregate_multiplication function. This is less complicated then one often thinks.
Oh and here is the solution using a recursive with clause. I build it upon nicosas version, but changed the math to be consistent with my previous examples.
Op wasn't very clear about the math behind and I think that led to much confusion.
with t as (
     select 1 slno, 1000 Amount , 0.12 Int_Rate from dual
     union all
     select 2 slno, 0 Amount , 0.13 Int_Rate from dual
     union all
     select 3 slno, 0 Amount , 0.14 Int_Rate from dual
  order by slno -- this is important
------ end of sample data ------
------ This removes the gap-problem, by introducing a sequential row number
, t2 as (select rownum rn, slno, Amount, Int_Rate
         from t)
------ trec below is a recursive with clause (AKA recursive subquery factoring) ------
trec(rn, slno,amount,int_rate) as
     select rn,slno,amount*(1+int_rate),int_rate
  from t2 base
  where rn=1
     union all
     select t3.rn,t3.slno,trec.amount*(1+t3.int_rate),t3.int_rate
     from t2 t3
     join trec     on t3.rn=trec.rn+1
------ just query the recursive with clause ------
select * from trec
RN SLNO AMOUNT INT_RATE
1    1   1120     0,12
2    2 1265,6     0,13
3    3 1442,784     0,14 This is the important line
trec.amount*(1+t3.int_rate)
Essentially this puts the amount from the previous row up for multiplication with interest from the current row. And sets it a the new amount.
Edited by: Sven W. on Aug 17, 2012 3:07 PM - added with clause solution
Edited by: Sven W. on Aug 17, 2012 3:10 PM - Typo correction