Reading a file and loading variables using function

Hi All, this is how my data looks like is there a way we can put these steps in a for loop or a function?

%Each .mat file looks like
A = 10; 
B = 30; 
C =15;
A_Threshold = 70; 
B_Threshold = 80; 
C_Threshold = 85;

%What I am interested in creating:
Day{1} = [10; %A
          30; %B
          15];%C
Day{2} = [20; %A
          40; %B
          10];%C
Day1Thrshld = [A_Threshold; B_Threshold; C_Threshold];
Day2Thrshld = [A_Threshold; B_Threshold; C_Threshold];

For me this data comes from a .mat file for each day. So each .mat file data will have all these variables as shown above

Day1.mat --> A,B,C and A_threshold, B_threshold, C_threshold

Day2.mat --> A,B,C and A_threshold, B_threshold, C_threshold

Day3..., Day4....

And I interested in appending/creating this Day{1},Day{2},Day{3},... variables with (A,B,C) values by reading the .mat files

The challenge for me in creating the function/for loop for above requirement is. The working directory is set but I am not sure how to pull a .mat file and then fetch these variables from it using a function/loop or the right approach.

I do know for a fact that timestamps would be used for different .mat files because let's day Day1.mat was created on 3/4/2023 and Day2.mat was created on 3/5/2023. However if the Day1 file (or any history file) is it is already used/fetched we do not want to use that/fetch data from.

Answer 1

[dpb]

A = 10;
B = 30;
C =15;
A_Threshold = 70;
B_Threshold = 80;
C_Threshold = 85;
save sample.mat A B C A_Threshold B_Threshold C_Threshold
whos -file sample

  Name             Size            Bytes  Class     Attributes

  A                1x1                 8  double              
  A_Threshold      1x1                 8  double              
  B                1x1                 8  double              
  B_Threshold      1x1                 8  double              
  C                1x1                 8  double              
  C_Threshold      1x1                 8  double              

S=load('sample.mat')

S = struct with fields:
              A: 10
    A_Threshold: 70
              B: 30
    B_Threshold: 80
              C: 15
    C_Threshold: 85

tSample=table([S.A;S.B;S.C],[S.A_Threshold;S.B_Threshold;S.C_Threshold],'VariableNames',{'Data','Threshold'});
day=ones(height(tSample),1);
tSample=addvars(tSample,day,'Before','Data','NewVariableNames',{'Day'})

tSample = 3×3 table
    Day    Data    Threshold
    ___    ____    _________

     1      10        70    
     1      30        80    
     1      15        85    

If the files all do use the same variable name for each, then you can just code them directly; and each is located positionally for each Day in the table. That may not be quite convenient-enough, you can also get the field names and use them as another variable for grouping or selection...

vars=fieldnames(S); vars=categorical(vars(1:2:end));
tSample=addvars(tSample,vars,'Before','Data','NewVariableNames',{'Unit'})

tSample = 3×4 table
    Day    Unit    Data    Threshold
    ___    ____    ____    _________

     1      A       10        70    
     1      B       30        80    
     1      C       15        85    

When you get the format for the timestamp, you can then parse it from the filename, converting it to a <datetime> variable, then test that it is/is not already included in the table before importing it.

This avoids the dreaded issue of having variable names that contain metadata like the UnitID or somesuch that make it extremely hard to write generic code without the use of the ugly eval technique.

You'll have to make real variable names for what it is these data represent; I just created a new data storage scheme that will be much simpler to use, making up names for the undefined...