sparse and non-integers

Hi all, I'm trying to create a sparse matrix using the sparse function: sp=sparse(i,j,s)

but the vector j that I'm using consists of double values. This leads to the error:

??? Error using ==> sparse Index into matrix must be an integer.

I don't want to round my j vector cause this would lead to a distortion of the results I need from my dataset.

Is there a way to overcome this problem and finally construct the sparse matrix i need using the original j vector?

Thanks in advance jaso

Answer 1

[Fangjun Jiang]

You don't have to modify j, just use

sp=sparse(i,round(j),s)

Answer 2

[Walter Roberson]

No, there is not. Matrices cannot be indexed at non-integral values.

What value would you want returned if they could be indexed at non-integral values? Should A(3.2) be (8/10 * A(3) + 2/10 * A(4)) for example?

You should perhaps be considering using something like ismember() to map your floating point values to arbitrary integers.