Col Sum function question

robertwf

I'm building a column formula in the formula editor and have run into a problem with the Col Sum function.

My dataset is structured as follows:

Claim a b
3154 1.5 10000
844 .5 1000000
7311 1.5 1000000
22777 .5 10000
. . .

and I'd like to sum the values according to the function:

Sum over i=1 to n {log(claim_i/b)*[(b/claim_i)^a + 1]^(-1)}.

The claim values are to be summed for all the rows in my dataset, hence a Col Sum function seems the natural choice. However, I don't want to sum the a and b variables by row - I want just the row value of each a and b to be included in the formula.

Normally I would use algebraic manipulations to bring the a and b constants outside of the summation. However, given the complexity of this formula, this isn't an option.

Is it possible to exclude column variables from the Col Sum function?

Many thanks,

Robert

dbailey

So does that make your formula is really

log(claim_i/b_i)*[(b_i/claim_i)^a_i + 1]^(-1)

if so, then could you use

sum(log(claim/b)*[(b/claim)^a + 1]^(-1))