How to extract values from structure as array and assign numeric values to them?

So i have a structure called data

looks like this

data. gender = 'female'

. age = '25'

.reproductiontime = 100x2 double

I would like to extract the values from age and gender from all my data which is around 200 files and they're all in the same format add a value to gender such as female = 1, male=2 non-binary=0

Can I assign values to gender structure or a sub-structure perhaps?

And do I just put this in the output of my function then? As in function [age, gender] = m(data) ?

Thanks in advance,

sorry if horribly basic but I feel like i'm ovecomplicating it and there might be a quick and easy answer I am not seeing.

[Turlough Hughes]

What size is the struct, and do you have code for loading each of the 200 files?

[Mau Dudas]

Well I tried to use

load ('repro_*.mat') which is the format name for all the 200 files

and then give

files= dir('repro_*.mat')

which is as far as i've got with it as it isn't quite right

the data structure for each repro_*.mat is a 1x1 with the abovementioned three fields.

Does that help?

Answer 1

[Ridwan Alam]

% assuming your 200 'data' structs are named as data001, data002, ...

varnames = whos('data*');
% since both 'age' and 'gender' properties are strings, string arrays are sufficient;
% can also use cell arrays
age = []; 
gender = [];
for v = 1:length(varnames)
    curData = eval([varnames(v).name]);
    age = [age; curData.age];
    % if you want to replace the gender string with numbers 1 and 2:
    gender = [gender; double(strcmp(curData.gender,'male'))+1];
    % else to just extract the gender string use this:
    % gender = [gender; curData.gender];
    
end

[Mau Dudas]

This looks pretty good. I'm gonna try it :)

[Mau Dudas]

Right, so what values would you assign to cur.data in this case? Cause I'm pretty sure that's what's missing

[Ridwan Alam]

varnames contain all the variables with names starting with data***

the assumption was that from the 200 files you have loaded the variables to your workspace as data001, data002, ... data200.

given your filename, this is the updated code:

filenames = dir('repro_*.mat');
% since both 'age' and 'gender' properties are strings, string arrays are sufficient;
% can also use cell arrays
age = []; 
gender = [];
for f = 1:length(filenames)
    data = load(filenames(f).name); 
    curData = data; 
    age = [age; curData.age];
    % if you want to replace the gender string with numbers 1 and 2:
    gender = [gender; double(strcmp(curData.gender,'male'))+1];
    % else to just extract the gender string use this:
    % gender = [gender; curData.gender];
    clear data;
end

Hope it works!

[Turlough Hughes]

If Mau has described the contents of the files repro_*.mat correctly then the above won't work as

data = load(filenames(f).name); 

is in fact nesting the 1x1 structure, stored in 'repro_*.mat' inside another structure called data, so in order to extract the data you would need replace:

age = [age; curData.age];
gender = [gender; double(strcmp(curData.gender,'male'))+1];

with

age = [age; curData.data.age];
gender = [gender; double(strcmp(curData.data.gender,'male'))+1];

If this did happen to work, it is because a given file repro_*.mat in fact contains the 3 variables, gender, age and reproductiontime.

Answer 2

[Turlough Hughes]

Assuming that when you load, the name of the variable is data as you mention in the question, then the following should do the job.

files = dir('repro_*.mat')
gender = cell(200,1);
age = cell(200,1);
gen_num = nan(200,1);

for c = 1:length(files)
    load(files(c).name);
    gender{c,1} = data.gender;
    gend_num(c,1) = strcmp(data.gender,'female')*1+strcmp(data.gender,'male')*2+strcmp(data.gender,'non-binary')*3;
    age{c,1} = data.age;
end

[Mau Dudas]

Just one more question.

Could you tell me what values belong to the 'cells' in the code?

As that is literally the only one I can't figure out...

Also it keeps asking to define 'data' for some reason

[Turlough Hughes]

Could you tell me what values belong to the 'cells' in the code?

Are you referring to these?

age = cell(200,1);
gender = cell(200,1);

What I do there is preallocate memory for storing the age and gender on each iteration of the for loop. I'm guessing you know this already but it's good practice to preallocate memory when you know the size the variables will be in advance.

Also it keeps asking to define 'data' for some reason

By this do you mean that you are getting the error:Undefined variable "data" or class "data.gender". If this is the case it is because the variables stored in each .mat file; 'repro_001.mat', 'repro_002.mat', etc, do not in fact contain a structure variable called data. You could try this:

fils=dir('repro_*.mat');
gender = cell(200,1);
age = cell(200,1);
gen_num = nan(200,1);
for c = 1:length(files)
    temp=load(fils(c).name);
    field_name=fields(temp);
    data=temp.(field_name{1,1});
    gender{c,1} = data.gender;
    gend_num(c,1) = strcmp(data.gender,'female')*1+strcmp(data.gender,'male')*2+strcmp(data.gender,'non-binary')*3;
    age{c,1} = data.age;
    clear data temp field_name
end

If that doesn't work, then write:

clear all
fils=dir('repro_*.mat');
load(fils(1).name);
whos

and let me know the result of whos.