[SOLVED ALREADY] Loop through data struct with variable names

EDIT:

Hi, So it seems like I was able to solve it already so yeah, Imma post the solution as an answer for people to look into it in case they have a similar problem.

Hi,

I am trying to loop through a structure called 'Erruptions'.

Now in this structure I have 3 more sub structs, thus it's length is 3.

The 3 structures have variable names 'E1', 'E2', etc. These are simply 1x2 doubles or vectors if you wanna call them that.

So by calling one of the structs I write

Erruptions.sub_struct

and I get its content

    E1: [6.0014 1.0161]
    E2: [4.5936 0.5191]
    E3: [3.2693 0.2849]
    E4: [2.7640 -0.0073]
    E5: [2.3443 0.1345]

But when I want to get the number of elements it says it's only 1, doesn't matter if I use

    length(Erruptions.sub_struct) % gives me ans = 1

or

    size(Erruptions.sub_struct) % gives me ans = [1 1]

which is very confusing. I want to run a loop through every substruct but since it only recognizes them as one single element, opposed to the above shown 5, itleave the substruct after taking the first into consideration.

How can I solve this problem?

[Stephen23]

"which is very confusing."

No, it is not.

The size of a structure is unrelated to how many fields it has.

"but since it only recognizes them as one single element, opposed to the above shown 5,"

A scalar structure only has one element, no matter how many fields it has. If you want to count the number of fields, why are you trying to get the size of the structure? Your confusion is caused by mixing up two different things.

"How can I solve this problem? "

Don't mix up the size of a structure with how many fields it has:

• If you want to loop over the elements of a structure then use its size/numel.
• If you want to loop over the fields of a structure then use its fieldnames.

[Sokratis Panagiotidis]

Hey, Now I uderstand my confusion. You were right about the mix up and the solution with 'fieldnames' is the correct one I was also able to solve.

But how would it confuse more people if I post the solution I found, since also someone else (besides you) was able to answer it this way?

[Steven Lord]

One analogy or mental model I think helps is that of a bookshelf. One bookshelf (struct) may have many shelves (fields) and each shelf may contain just one item or a series of items (data in the field.)

Answer 1

[Voss]

% creating a structure like yours
Erruptions = struct('sub_struct',struct( ...
    'E1',[6.0014 1.0161], ...
    'E2',[4.5936 0.5191], ...
    'E3',[3.2693 0.2849], ...
    'E4',[2.7640 -0.0073], ...
    'E5',[2.3443 0.1345]))

Erruptions = struct with fields:
    sub_struct: [1×1 struct]

% Erruptions.sub_struct is a scalar structure, so its length is 1
Erruptions.sub_struct

ans = struct with fields:
    E1: [6.0014 1.0161]
    E2: [4.5936 0.5191]
    E3: [3.2693 0.2849]
    E4: [2.7640 -0.0073]
    E5: [2.3443 0.1345]

% you want to get the fields of Erruptions.sub_struct
fn = fieldnames(Erruptions.sub_struct)

fn = 5×1 cell array
    {'E1'}
    {'E2'}
    {'E3'}
    {'E4'}
    {'E5'}

% and loop through those
for ii = 1:numel(fn)
    Erruptions.sub_struct.(fn{ii})
end

ans = 1×2

6.0014 1.0161

ans = 1×2

4.5936 0.5191

ans = 1×2

3.2693 0.2849

ans = 1×2

2.7640 -0.0073

ans = 1×2

2.3443 0.1345

[Sokratis Panagiotidis]

Yeah this was pretty much the solution I was able to find only two minutes after posting that question except for using 'length' instead of 'numel'.

Thanks !

[Voss]

You're welcome!

Answer 2

[Benjamin Thompson]

You want to index into the structure array, then pass one of those array elements to the length function:

>> S1.E1 = [1 1];

>> S2.E1 = [1 1];

>> SArray(1) = S1

SArray =

struct with fields:

E1: [1 1]

>> SArray(2) = S2

SArray =

1×2 struct array with fields:

E1

>> length(SArray.E1)

Error using length

Too many input arguments.

>> length(SArray(1).E1)

ans =

2

[Stephen23]

A much simpler way to create that structure:

S(1).E1 = [1,1];
S(2).E1 = [0,1];

Answer 3

[Sokratis Panagiotidis]

SOLUTION:

Just like stated above I was able to solve it so here's ONE solution I managed to find.

What I did is I took the names of the struct, meaning

    fn = fieldnames(struct);

and in the next step created a for loop

    for j = 1:length(fn{j})

and THEN I did the exact same thing and took the fieldnames of each substruct, meaning

        newfn = fieldnames(struct.(fn{j}));

which is

    E1: [6.0014 1.0161]
    E2: [4.5936 0.5191]
    E3: [3.2693 0.2849]
    E4: [2.7640 -0.0073]
    E5: [2.3443 0.1345]

Now I try to get its length and voilà! I can use it in aother for loop to continue:

    for i = 1:length(newfn)

So yeah, there you have it!