Multiplication of sparse matrix and a normal matrix

I have a Sparse Matrix : S

The size of S is : 196608x4096 sparse double

and the normal (Non Sparse) matrix : D

The size of D is : 196608x1 single

Matrix D is full of one . [ means : D= ones(196608,1 )]

I want to multiply them both >> transpose(S)* D;

But showing me some error :

1. " Error using .* Sparse single array arithmetic operations are not supported."
2. " Error using * MTIMES (*) is not supported for one sparse argument and one single argument "

Even I tried to convert the matrix D into sparse but showing me following error :

" Error using sparse Input matrix must be double or logical".

Any kind of suggestions to solve this problem will be appreciated.

Thank You !!!

Happy Weekend !!!

Answer 1

[Walter Roberson]

You are trying to multiply a sparse double precision matrix by a non-sparse single precision matrix.

If you are expecting:

• sparse double result: use S.' * double(D)
• sparse single result: not permitted. sparse matrices can only be double precision or logical
• full double result: use full(S.' * double(D))
• full single result: use full(S.' * D)

Answer 2

[David Goodmanson]

Hi Amit,

sum(S)'

where the result is still sparse, or

full(sum(S)')

if you want a regular vector at that point.

[Amit Chakraborty]

I do not want to convert into "full". Because above matrix is just example. Converting into "full" will takes lot of memory.

Answer 3

[Amit Chakraborty]

@Walter Roberson , @David Goodmanson

Thank you Both. I have learnt a new thing from you guys.