Defining a 4d double

I am trying to build a 4D matrix from a velocity time series so that I can use it as an input to another program. My wind field is a M*N matrix where M is the number of points on the square grid and N is the measurement time in seconds (sampling at 1 hz). I now want to build a 4D array following which should look like this

4D_matrix=[time(1*N) velocity(M*N) x_coord(M,1) y_coord(M,1)];

How do i write this?

Thank you, Ananth

[Rik]

Have you looked into reshape?

If that doesn't help you, what dimensions should this 4D matrix have? (N,M*N,M,M)? You can generate an empty matrix with these dimensions with zeros(N,M*N,M,M) or ones(N,M*N,M,M) and then fill it with a loop if you can't make it work any other way.

[Anantha Padmanabhan]

Hello, I have tried this already but I get an error that my array size exceeds maximum array size preference as the number of points M=676 and time N=600. Is there a way around this?

[KSSV]

What could be the sizes of M and N? If you create 4D matrix, you may end up with memory.

[Anantha Padmanabhan]

As said, M=676 (number of grid points) N=600 (for 10 minute simulations)

[Faqih Romi]

d=[2 3 4 1; 3 2 1 4];

dd(:,:,:,1)=d(1,:,1);

dd(:,:,:,2)=d(2,:,1);

Answer 1

[Guillaume]

Well, you need 676^3 * 600^2 * 8 bytes ~= 890 TB of memory to store that many elements, regardless of the shape. Chances are you don't have a computer with that much memory (not even in the ballpark).

You need to rethink what you want to do. In particular, I would think that x_coord and y_coord correspond to the rows and columns of your velocity field, therefore it does not make sense to store them in additional dimensions of the same matrix (as that would wastefully repeat the same values).

Answer 2

[Pawel Tokarczuk]

I think what you need is something like:

Velocity = zeros([NROWS, NCOLS, NTIMES], 'double');

[ X, Y ] = meshgrid(0:NCOLS-1, 0:NROWS-1);

XX = X0 + ScaleX*X;
YY = Y0 + ScaleY*Y;

T  = double(0:NTIMES-1);
TT = T0 + ScaleT*Time;

Note the order of the indices (column-major in the first two dimensions, last index varies most slowly).

Those should be the basic data structures; now, you just need to use them appropriately.

IHTH