How to Put These Vectors in a Matrix?

A.B.G = [1 2 3];
A.C.G = [4 5 6];
A.D.G = [7 8 9];
A.E.G = [10 11 12];
A.F.G = [13 14 15];

so it will look like:

1  2  3
4  5  6
7  8  9
10 11 12
13 14 15

It's just an example and A has more than 5 fields, so please do not suggest something like this:

[A.B.G; A.C.G; A.D.G; A.E.G; A.F.G]

Answer 1

[Stephen23]

Here is one way:

>> A.B.G = [1 2 3];
>> A.C.G = [4 5 6];
>> A.D.G = [7 8 9];
>> A.E.G = [10 11 12];
>> A.F.G = [13 14 15];
>> cell2mat(structfun(@(s){s.G},A))
ans =
     1     2     3
     4     5     6
     7     8     9
    10    11    12
    13    14    15

Although to be honest your code would be much faster and simpler if you used a non-scalar structure, because then you can simply do this:

>> S(1).G = [1 2 3];
>> S(2).G = [4 5 6];
>> S(3).G = [7 8 9];
>> S(4).G = [10 11 12];
>> S(5).G = [13 14 15];
>> vertcat(S.G)
ans =
   1     2     3
   4     5     6
   7     8     9
  10    11    12
  13    14    15

Using indices will be much simpler than trying to access fieldnames, and woudl allow you to very simply solve your other question by using indices to select the parts of the structur that you want:

>> vertcat(S([1,3,4]).G)
ans =
     1     2     3
     7     8     9
    10    11    12

Do not make your code more complicated than it needs to be. Good data storage makes your code simpler and neater too.

[Stephen23]

Hmm, just noticed this identical question and answer:

https://www.mathworks.com/matlabcentral/answers/328109-how-to-use-a-for-loop-to-dynamically-assign-the-values-of-struct-fields-to-a-matrix

[Guillaume]

"Although to be honest your code would be much faster and simpler if you used a non-scalar structure"

Actually, judging by previous questions, the code would be much faster and simpler if a structure wasn't used at all. The fields don't appear to have any purpose other than generic containers. For this, cell arrays or matrices would be much better.

[Rightia Rollmann]

Thank you so much for your patience and brilliance!

Your suggestion of using non-scalar structure is generally great! Specifically speaking for this data set, I thought if I name each field separately, it would be easier to recognize them later by their field names

[Stephen23]

@Rightia Rollmann: it is better to store meta-data as data in its own right, e.g.:

 S(1).data = [1,2,3]; % measurements
 S(1).name = 'anna';
 S(2).data = [4,5,6];
 S(2).name = 'bob';

rather than trying to make the fieldnames equal to the subjects names. This is simpler and easier to write code for, which means your code will be faster and less buggy.