Apply logical mask to every matrix in array

I have a 3D array of size 640x640x20 i.e. 20 matrices of size 640x640. In addition I have a logical mask of size 640x640. How can I apply this mask to every matrix in the array? Is this only possible with a for loop?

for jj=1:20
    mat_temp=B(:,:,jj);
    mat_temp(mask==1)=NaN
    B(:,:,jj)=mat_temp;
end

%% I tried the following but it didn't worked (mask was only applied to first matrix)
B(mask==1)=NaN;

Answer 1

[Jan]

B    = reshape(1:24, 2,3,4);
mask = logical([1,0,1; 0,1,0]);

sB = size(B);
B  = reshape(B, [], sB(3));  % Join the first two dimensions
B(mask, :) = NaN;
B  = reshape(B, sB)

D = 
D(:,:,1) =

   NaN     3   NaN
     2   NaN     6


D(:,:,2) =

   NaN     9   NaN
     8   NaN    12


D(:,:,3) =

   NaN    15   NaN
    14   NaN    18


D(:,:,4) =

   NaN    21   NaN
    20   NaN    24

% Alternatively:
M = ones(size(mask));
M(mask) = NaN;
B = B .* M;

Note: No need to compare mask with 1: mask==1. Use mask directly, if it is a logical array.

[Jan]

A speed coparison with R2018b:

X    = rand(640, 640, 20);
mask = rand(640, 640) > 0.6;

rep = 1e2;

B = X;
tic;
for k = 1:rep
   for jj=1:20
      mat_temp          = B(:,:,jj);
      mat_temp(mask==1) = NaN;
      B(:,:,jj)         = mat_temp;
   end
end
toc  % Elapsed time is 8.416893 seconds.

B = X;
tic;
for k = 1:rep
   sB = size(B);
   B  = reshape(B, [], sB(3));
   B(mask, :) = NaN;
   B = reshape(B, sB);
end
toc  % Elapsed time is 0.966218 seconds.

B = X;
tic;
for k = 1:rep
   M = ones(size(mask));
   M(mask) = NaN;
   B = B .* M;
end
toc  % Elapsed time is 1.173872 seconds.

B = X;
tic;
for k = 1:rep
   s    = size(B);
   f    = find(mask==1) + (0:s(3)-1) .* (s(1)*s(2));
   B(f) = nan;
end
toc  % Elapsed time is 2.992846 seconds.

[Marcel345614]

Thank you! This was helpful!

Answer 2

[David Hill]

s=size(B);
f=find(mask==1)+(0:s(3)-1).*(s(1)*s(2));
B(f)=nan;

Answer 3

[Loïc Reymond]

The most compact way would probably be:

bsxfun(@(x,y) x.*y,mat,mask)

Or alternatively (probably slower):

mat.*repmat(mask,1,1,size(mat,3))

Where mat is your 640x640x20 matrix and mask your masking array.