How to fill a matrix in a timeseries with one single time step?

Hello,

I want to fill a 2x3 matrix in a timeseries for simulating in simulink with one single time step. It works if I have more than one time step: The series.Data gets the size 2x3x[NumberOfTimesteps].

But when I only have one time step it does not work to use the reshape command because a 2x3x1 matrix results to a 2x3 matrix. Then simulink interpretates the matrix as a 1x3 array.

Is there any solution? I think it would work, if I could force a 2x3 matrix to a 2x3x1 matrix.

Edit: I want to use the timeseries to load the data in the 'FromWorkspace' Block.

[Walter Roberson]

Every 2 x 3 matrix is also a 2 x 3 x 1 matrix.

A = rand(2,3)

A = 2×3

0.3466 0.1710 0.4546 0.1404 0.3761 0.1906

for K = 1 : 5
   fprintf('dimension %d --> %d\n', K, size(A,K));
end

dimension 1 --> 2
dimension 2 --> 3
dimension 3 --> 1
dimension 4 --> 1
dimension 5 --> 1

[Simulant]

Thanks for your answer! :-)

But that does not help me in the simulink model. The data is only interpretated correctly if the simulation has more than one time step and the timeseries.Data has a size of [2, 3, length(time)], if length(time)>1.

[Walter Roberson]

It is not possible to create a 2 x 3 x 1 array that can be detected as being different than a 2 x 3 array with the same elements.

[Simulant]

Thanks for helping me! :-)

I found a simple solution to avoid the problem:

I added a dummy step to my simulation after the first step, to get the simulation running. From the simulation result I remove the last dummy step.

But still it would be nice, if Matlab could see a difference between a [x, y] and a [x, y, 1] matrix.