Read multiple files and store fitting parameters & graphs

I have a function that reads from an input file (floating point numbers, only) columns (A, B, C, D, E, F, G, H, I) perform some mathematical operations and plot some fittings.

Since I have multiple input files named as: res1, res8, ..., resn, attention not in sequencial order (meaning not having res2, for example) how should I do for my script read all those input files and for each run/analysis create an output file results1, results8 ... etc, with fitting results and plots?

Any help would be much appreciated.

[Stephen23]

Use the dir method:

https://www.mathworks.com/help/matlab/import_export/process-a-sequence-of-files.html

[Angelo Figueiredo]

Still, the outputs aren't save in the local directory?!

Besides, I don't understand why I only get one fitting result for the two *.txt input files? That apparently belongs to the last readed *.txt file

function [results] = test(A,B,C,D,E,F,G,H,time)

numfiles = 2;
mydata = cell(1, numfiles);

for k = 1:numfiles
  myfilename = sprintf('Res%d.txt', k);
  mydata{k} = importdata(myfilename);
end


A = mydata{k}(:,1); 
B = mydata{k}(:,2); 
C = mydata{k}(:,3);
D = mydata{k}(:,4); 
E = mydata{k}(:,5);
F = mydata{k}(:,6); 
G = mydata{k}(:,7);
H = mydata{k}(:,8);
time = mydata{k}(:,9);

ABsub = E - E(1);
BAsub = G - G(1);
ref = A(1) + C(1);
AA = A/ref;
BB = B/ref;
AB = ABsub/ref;
BA = BAsub/ref;

[Bob Thompson]

You're only getting one set of outputs because you're only looking at mydata{k}, where k is numfiles (the last one). If you want to perform the operation for all of the files, move them all inside the loop, including the output.

[Angelo Figueiredo]

Sorry would be possible to give an example.

Answer 1

[Bob Thompson]

function [results] = test(A,B,C,D,E,F,G,H,time)

numfiles = 2;
mydata = cell(1, numfiles);

for k = 1:numfiles
  myfilename = sprintf('Res%d.txt', k);
  mydata{k} = importdata(myfilename);
  A = mydata{k}(:,1); 
  B = mydata{k}(:,2); 
  C = mydata{k}(:,3);
  D = mydata{k}(:,4); 
  E = mydata{k}(:,5);
  F = mydata{k}(:,6); 
  G = mydata{k}(:,7);
  H = mydata{k}(:,8);
  time = mydata{k}(:,9);

  ABsub = E - E(1);
  BAsub = G - G(1);
  ref = A(1) + C(1);
  AA = A/ref;
  BB = B/ref;
  AB = ABsub/ref;
  BA = BAsub/ref;
  % Insert new database creation here
  % Insert output command here
end

Answer 2

[Angelo Figueiredo]

Sorry once again but I still don't understand why I am getting one plot from my dataset. Besides how can I create at the end of the running script an output file with all fitting results for all files executed

function [results] = test(A,B,C,D,E,F,G,H,time)
numfiles = 2;
mydata = cell(1, numfiles);
for k = 1:numfiles
  myfilename = sprintf('Res%d.txt', k);
  mydata{k} = importdata(myfilename);
  A = mydata{k}(:,1); 
  B = mydata{k}(:,2); 
  C = mydata{k}(:,3);
  D = mydata{k}(:,4); 
  E = mydata{k}(:,5);
  F = mydata{k}(:,6); 
  G = mydata{k}(:,7);
  H = mydata{k}(:,8);
  time = mydata{k}(:,9);
  ABsub = E - E(1);
  BAsub = G - G(1);
  ref = A(1) + C(1);
  AA = A/ref;
  BB = B/ref;
  AB = ABsub/ref;
  BA = BAsub/ref;
  % Insert new database creation here
  % Insert output command here

[AAcol,AArow] = size(normAA);
if AArow > AAcol
    normAA = normAA';
end
[BBcol,BBrow] = size(normBB);
if BBrow > BBcol
    normBB = normBB';
end

end

end

%estimated values
default.k = 3.2;
default.pA = 0.5;
default = [default.k;default.pA];

[fitparameters] = fminsearch(@(x) zex(x,normAA,normBB,time),default);
k = fitparameters(1);
pA = fitparameters(2);

[residual,IAA,IBB] = zex(fitparameters,normAA,normBB,time);
results.k = k;
results.pA = pA;
results.datafit = [normAA normBB];
results.dataIAA = normAA;
results.dataIBB = normBB;
results.fitIAA = IAA;
results.fitIBB = IBB;

figure
hold on
plot(time,normAA,'o','color','r')
box on
fprintf('\n~~~~Results~~~~\nkex: %f\npA: %f\npB: %f\nR1A: %f\nresidual: %f\n',k,pA,residual)
end

function [residual,IAA,IBB] = zex(default,Iaa,Ibb,time)
k = default(1);
pA = default(2);
number_times = length(time);

aAA = zeros(number_times,1);
aBB = zeros(number_times,1);

for n = 1:1:number_times
    aAA(n) = pA*(pA+pB*exp(-kex*time(n)))*exp(-R1A*time(n));
end

end

[Bob Thompson]

Move everything except for the zex function inside the loop.

[Angelo Figueiredo]

Done that before but I get an error:

"Field assignment to a non-structure array object.

Error in k (line 70)

default.k = 3.2;"

[Bob Thompson]

You are receiving that error because your structure field is named the same as your index (k). I would recommend changing the name of the structure field, since it should be a unique variable name anyway.

[Angelo Figueiredo]

Thanks a lot, Bob! But could you exemplify briefly, I am struggling to understand what you mean by structure field

[Bob Thompson]

To the best of my knowledge, almost any time a variable is specified using dot convention it is considered a structure by default. You are defining 'default.k' as equal to a value. In this command 'default' is the structure name, and 'k' is the name of the field in the structure. You cannot do this within your for loop because you have 'k' assigned as the index variable of the loop. This can be simply solved by changing default.k into something like default.kVal, or whatever you would like to name it as. (Subsequent callings of default.k will also need to be changed.)

[Angelo Figueiredo]

Seems not easy, unfortunately.

I've done as you said but still I get the same error message

"Field assignment to a non-structure array object.

Error in k (line 70)

default.ex = 3.2;"

[Bob Thompson]

Mmm, I forgot this part. Put brackets around your new structure field definitions.

[default.ex] = 3.2;
[default.pA] = ...

[Angelo Figueiredo]

Hi again, still I don't understand why I only get one plot and not two plots as read by 'numfiles', as well as only one fitting result output instead of two?