For loop to save doubles in .mat files

I have a .mat file which stores the variables:

E.g: name: ABC, type : 20 x 1; class: double

name: DEF, type : 20 x 1; class: double

name: GHI, type : 20 x 1; class: double

name: JKL, type: 1 x 1; class: struct

name: JKL, type: 1 x 1; class: struct

name: XYZ, type: 7 x 1; class: cell

name: LMN, type : 30 x 1; class: double

etc.

I would like to use only the (20 X1) doubles one by one in another line of code. For instance,

load_mat = load('some_mat_file.mat')
new_var = load_mat.DEF;
new_var = load_mat.GHI;

But instead of writing repeatedly, could you provide an example of a for loop or something else? Thanks a lot in advance.

[Adam Danz]

Why do you need to remove the data from the neatly organized structure (load_mat)?

Just use load_mat.DEF instead of new_var.

[Stephen23]

"could you provide an example of a for loop..."

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

"... or something else?"

https://www.mathworks.com/matlabcentral/fileexchange/26216-structure-fields-to-variables

Answer 1

[Voss]

load_mat = load('some_mat_file.mat')

load_mat = struct with fields:
    ABC: [20×1 double]
    DEF: [20×1 double]
    GHI: [20×1 double]
    JKL: [1×1 struct]
    LMN: [30×1 double]
    XYZ: {7×1 cell}

% create a new structure, "new_mat", which contains 
% only the 20x1 doubles from load_mat:
new_mat = struct();
f = fieldnames(load_mat);
for ii = 1:numel(f)
    if isa(load_mat.(f{ii}),'double') && isequal(size(load_mat.(f{ii})),[20 1])
        new_mat.(f{ii}) = load_mat.(f{ii});
    end
end

new_mat

new_mat = struct with fields:
    ABC: [20×1 double]
    DEF: [20×1 double]
    GHI: [20×1 double]

[alphabetagamma]

This is very helpful. Thank you

[Voss]

You're welcome!

[alphabetagamma]

Actually, I have one last follow-up question. Let's say that from the new_mat structure, I want to only save the first two doubles, how I do that? More generally, if there are n doubles, and I would like to store( n-p) doubles, then what would be a good approach? I tried the following but it didn't work:

Perhaps it works on for other types such as in matrices.

new_mat1 = new_mat(1:end -1)

[Voss]

load_mat = load('some_mat_file.mat')

load_mat = struct with fields:
    ABC: [20×1 double]
    DEF: [20×1 double]
    GHI: [20×1 double]
    JKL: [1×1 struct]
    LMN: [30×1 double]
    XYZ: {7×1 cell}

N_keep = 2; % keep the first at-most-2 20x1 doubles
keep_count = 0;

new_mat = struct();
f = fieldnames(load_mat);
for ii = 1:numel(f)
    if isa(load_mat.(f{ii}),'double') && isequal(size(load_mat.(f{ii})),[20 1])
        new_mat.(f{ii}) = load_mat.(f{ii});
        keep_count = keep_count+1;
    end
    if keep_count == N_keep
        break
    end
end

new_mat

new_mat = struct with fields:
    ABC: [20×1 double]
    DEF: [20×1 double]

Note that the order of the fields will be as returned by fieldnames.

[alphabetagamma]

Sorry I forgot to reply earlier. Thanks a ton again :)

[Voss]

You're welcome!