split mtatrix and name automatically

Hello guys

I have a mat file (attached) which is a 10004*15 matrix. I need to split it into 15 seprated matrixes like (10004*1). and I need to name these matrixes like ur1, ur2, ...ur15.

I know that ı can split matrixes by using the following code:

ur1=ur(:,15);
ur2=ur(:,14)
% and so on....

but I want to know if I can use it in a for loop so that I can have fewer lines (since I need to use the same code for other data files and it will be so long).

Also I want to ask if there is a way to name my output matrixes automatically.

[Walter Roberson]

Yes it is possible to do that with a for loop, naming the arrays the way you ask. However we need to ask why you are doing that? What reason do you have for splitting into separate variables instead of just indexing the single variable?

[mehra]

the reason I need to split is that later I use a function which uses (n,1) matrixes, since my matrix is (n,15) I need to split it.

[Stephen23]

"the reason I need to split is that later I use a function which uses (n,1) matrixes, since my matrix is (n,15) I need to split it."

Nothing in your explanation requires dynamically named variables. Your suggested approach would be slow, complex, inefficient, insecure, and obfuscate the code intent:

https://www.mathworks.com/matlabcentral/answers/304528-tutorial-why-variables-should-not-be-named-dynamically-eval

Everything you have explained so far would be simpler and much more efficient using standard approaches e.g. indexing into a cell array or 3D array. Your question is a good example of this:

https://xyproblem.info/

https://en.wikipedia.org/wiki/XY_problem

[mehra]

As I mentioned my goal was to see by using (n,1) matrix in my function the results would be better despiked. I needed to try both to compare the results.

thanks

[Stephen23]

"... I need to name these matrixes like ur1, ur2, ...ur15."

If the code requires lots of numbered variable names then you are doing something wrong.

Using indexing is simpler, neater, easier, and much more efficient than your approach. What is stopping you from using the indexing directly, or a simple cell array?

Answer 1

[the cyclist]

Anywhere that you could have used the variable name ur1, you can almost always just reference ur(:,1) directly instead, and not used the dynamically named variables. (Please read this post about why dynamically named variables are a terrible idea in general.)

If you must break out into individual variables, then using a cell array is typically a better idea:

% Initialize empty cell array
urc = cell(15,1);

% Fill them
for nc = 1:15
    urc{nc} = ur(:,nc); % Note the use of curly brackets to denote contents of cell. See docs.
end

[mehra]

Thank you ,

I also rather not using dynamically named variable but previously I used it in that way my despiking function but apparently it did not despiked and my values were still included outliers. It ıs mentioned in the guidance of ''func_despike_phasespace3d'' that fi in [fo, ip] = func_despike_phasespace3d( fi, i_plot, i_opt ) should be (n,1) matrix. So I thought it maybe better to split the matrixes and use them as one-dimensional matrix.

My previous code is like:

for np=1:15;
    ur(:,np)=Data.Profiles_VelX(:,np);
    vr(:,np)=Data.Profiles_VelY(:,np);
    w1r(:,np)=Data.Profiles_VelZ1(:,np);
    w2r(:,np)=Data.Profiles_VelZ2(:,np);
    [u(:,np), ip1] = func_despike_phasespace3d( ur(:,np),8, 2); %here I have directly refered to the matrix
    [v(:,np), ip2] = func_despike_phasespace3d( vr(:,np),8, 2);
    [w1(:,np), ip3] = func_despike_phasespace3d( w1r(:,np), 8, 2);
    [w2(:,np), ip4] = func_despike_phasespace3d( w2r(:,np), 8, 2);
    w(:,np)=(w1(:,np)+w2(:,np))/2;
end