Replace for loop for cell assignment based on anonymous function

I have a anonymous function of x which outputs a matrix. I have an array of 'x' values for which I want to store output of anonymous function in individual cell using for loop. I was wondering if this for loop can be replaced for better execution efficiency. Sample code below:

anonfunc = @(x) [zeros(2); blkdiag(1-x, 1+x)];
input = [1,5,7]';
output = cell(1,3);
for i = 1:3
output{i} = anonfunc(input(i));
end

I want to replace for loop with single command such as:

output{1:3} = anonfunc(input); %I am well aware this command doesn't work and is written to give idea of what I want.

Thanks in Advance.

Answer 1

[csamoa]

You can use arrayfun with 'UniformOutput' set to false, which will output a cell-array of the results of the anonfunc on the values of your input vector.

The orientation of the output cell-vector depends on the orientation of the input vector.

The optional transpose changes the orientation.

anonfunc = @(x) [zeros(2); blkdiag(1-x, 1+x)];
input = [1,5,7]';

tic
output = cell(1,3);
for i = 1:3
output{i} = anonfunc(input(i));
end
toc

tic
output2 = arrayfun(anonfunc, input, 'UniformOutput', false)';
toc

The gain in speed varied, e.g.:

Elapsed time is 0.003087 seconds.

Elapsed time is 0.001898 seconds.

[Nauman Haider]

Thanks. The suggestion does output what I was looking for. However, I was expecting much faster performance in my code but it was not as much as I thought it will be. Thanks once again for your help @csamoa.