How to avoid writing a for loop in A\b when A and b are different dimensions

I am currently solving where A is a 3d complex array of dimension and b is a 2d real array of dimension . In order to do this, I create a single for loop which calls linsolve() times.

Could I avoid writing this for loop so that I wouldn't have to spend the extra computation calling linsolve() times? It appears that the pagemldivide() function might be one way of doing this.

clear all; close all;

% Original Method 

Nx = 32;
A = complex(rand(Nx+1,Nx+1,Nx),rand(Nx+1,Nx+1,Nx)); % A is a 3D complex array
b = rand(Nx+1,Nx); % b is a 2D real array
soln = zeros(Nx,Nx+1); % the resulting solution will be a 2d complex array

for j=1:Nx
    prob = linsolve(A(:,:,j),b(:,j)); % A\b
    soln(j,:) = prob.'; % Size (Nx,Nx+1)
end

% New Method

soln_tilde = pagemldivide(A(:,:).',b(:)); % Size (Nx+1,1)

Is there a way of using pagemldivide() (or a seperate Matlab function) to avoid writing a for loop and get the same resulting solution of size ?

Answer 1

[Matt J]

soln=pagemldivide(A,reshape(b,[],1,Nx));

soln=soln(:,:).';

[Matthew Kehoe]

Wow! I think there is an extra comma after the second colon. Auto-reshaping b to the same size as A through is an amazing Matlab feature.

[Matt J]

Wow! I think there is an extra comma after the second colon.

There was, but I fixed it. It should work now. If it does, please Accept-click the answer.