Indexing one particular dimension regardless of number of dimensions

I have a structure of several dimension (7 right now) that I will be adding dimensions to in the future. In a section of my code I need to isolate the third dimension into it's three components which right now I do like

outX = in(:,:,1,:,:,:,:); outY = in(:,:,2,:,:,:,:); outZ = in(:,:,3,:,:,:,:);

Is there a way to generalize this without having to edit the number of colons in the statement?

ex:

outX = coolFunction(in,3,1); outY = coolFunction(in,3,2); outZ = coolFunction(in,3,3);

Thanks!

Answer 1

[Guillaume]

function out = coolFunction(in, dim, page)
   validateattributes(dim, {'numeric'}, {'scalar', 'positive', '<=', ndims(in)});
   validateattributes(page, {'numeric'}, {'scalar', 'positive', '<=', size(in, dim)});
   alldims = arrayfun(@(d) 1:d, size(in), 'UniformOutput', false);
   alldims{dim} = page;
   out = in(alldims{:});
end

[Jacob Matthews]

Super cool answer... and I won't pretend to understand it initially, but I'll study it.

I'm mostly glad I wasn't overlooking some existing function or indexing notation.

Thanks!

[Guillaume]

Basically, it creates a cell array (with arrayfun) where each cell is an array from 1 to the size of each dimension (what colon would do), then replace the cell of the required dimension with just the page number you want in that dimension.

The last line uses expansion of cell array into comma separated list to index the matrix.