Reshape 1D to 3D array

I have a 1D array (365x1) that I need to reshape into a 3D array (365x1x1) to divide into 3D array (365x721x1440). What's the most efficient means of accomplishing this task?

[Torsten]

I don't understand your task.

Answer 1

[the cyclist]

MATLAB arrays implicitly have length-1 dimensions after the defined dimensions. For example

% Define 365x1 array
M = rand(365,1);

% Length in 3rd dimension
size(M,3)

ans = 1

% Length in 31st dimension
size(M,31)

ans = 1

Regarding how that ends up being (365x721x1440), you'll need to provide more detail.

[John Cruce]

Let me try again.

A=rand(365,721,1440);
B=[1:365]'

I want to divide A by B along the first dimension (365). My thought was that I'll need to reshape B then divide the two. Does that sound right?

[Torsten]

You mean

A=rand(365,721,1440);
B=[1:365]';
for i = 1:365
    A(i,:,:) = A(i,:,:)/B(i);
end
A

?

[DGM]

I think we're getting confused by the suggestion of "reshaping". The way it appears to me, reshaping A is not necessary. It appears as this is simply a misunderstanding how to make array sizes compatible for arithmetic operations. Consider the example:

A = (1:4).' % a column vector

A = 4×1

1 2 3 4

B = 2*ones(4,4,2) % a 3D array with the same page height

B = 
B(:,:,1) =

     2     2     2     2
     2     2     2     2
     2     2     2     2
     2     2     2     2


B(:,:,2) =

     2     2     2     2
     2     2     2     2
     2     2     2     2
     2     2     2     2

% you could expand A to have the same size as B
% A = repmat(A,[1 4 2]);

% but it's not necessary
A./B

ans = 
ans(:,:,1) =

    0.5000    0.5000    0.5000    0.5000
    1.0000    1.0000    1.0000    1.0000
    1.5000    1.5000    1.5000    1.5000
    2.0000    2.0000    2.0000    2.0000


ans(:,:,2) =

    0.5000    0.5000    0.5000    0.5000
    1.0000    1.0000    1.0000    1.0000
    1.5000    1.5000    1.5000    1.5000
    2.0000    2.0000    2.0000    2.0000

So long as the arrays are compatibly-sized along their non-singleton dimensions, you should be able to perform elementwise arithmetic without explicitly expansion. If you were running a version older than R2016b, the recommended method would be to use bsxfun() instead of repmat().

Of course, I might be interpreting the question wrong.

[John Cruce]

You're excatly right. I was definitely overcomplicating it.