how to get back to the main Function after a for loop is finished

A function where I have a FOR loop runs indefinitely when that function is a sub function of the main function. How can I get back to the main function when the FOR loop is done? this problem happends when the input data is read through a text file otherwise the code runs fine. This is the code, the first part is the main function and the second function is the sub function. I have attached the text file as well. Thanks.

function [x,v,SHmaxx]=MySHmaxMatrixPolygon(x1)


%input data
load PolyDa.txt;

miu=0.6;
pp=PolyDa(1,8);
S2=PolyDa(1,10);
S3=PolyDa(1,11);  
Sv=S2;


%Polygon boundaries

Fmiu=(sqrt(miu^2+1)+miu)^2;
A1=((Sv-pp)/Fmiu)+pp;
D1=Fmiu*(Sv-pp)+pp;
n=round(D1-A1)-1;

A2=((Sv-pp)/Fmiu)+pp;
B1=((Sv-pp)/Fmiu)+pp;
B2=Sv;
C1=Sv;
C2=Fmiu*(Sv-pp)+pp;

D2=Fmiu*(Sv-pp)+pp;
E1=Sv;
E2=Sv;
F1=A1;
F2=A2;
xx = [A1 B1 C1 D1 E1 F1 B1 E1 E1 C1];
yy = [A2 B2 C2 D2 E2 F2 B2 E2 E2 C2];


%Calling function polygon to calculate SHmax(x) at UCS1

z=zeros(n,1)+x1;

[x,fval]=fsolve(@polygon,z);
end


%% THE SUB FUNCTION THAT HAS THE ISSUE WITH THE FOR LOOP IS THIS ONE:

function P=polygon(xi)
% data

load PolyDa.txt;


    trend=PolyDa(1,4);
    plunge=PolyDa(1,5);
    rake=PolyDa(1,6);
    nu=PolyDa(1,13);
    pp=PolyDa(1,8);
    Pm=PolyDa(1,12);
   
    S2=PolyDa(1,10);
    
    wbo=PolyDa(1,7);

    FA=PolyDa(1,15);
    Biot=PolyDa(1,14);
    azi=PolyDa(1,3);
    incl=PolyDa(1,2);
    To=PolyDa(1,9);
    depth=PolyDa(1,1);

Sv=S2;
miu=0.6;

Fmiu=(sqrt(miu^2+1)+miu)^2;
A1=((Sv-pp)/Fmiu)+pp;
D1=Fmiu*(Sv-pp)+pp;
C1=Sv;


n=round(D1-A1)-1;
Sx=zeros(n,1);
S3=zeros(n,1);
Sgf=zeros(3,3);
Rbf=zeros(3,3);
Sbf=zeros(3,3);
Sf=zeros(3,3);
thetamax=zeros(n,1);
sTZf=zeros(n,1);
sZZf=zeros(n,1);
sTTf=zeros(n,1);
thetab=zeros(n,1);
st_min=zeros(n,1);
st_max=zeros(n,1);
Sigma1=zeros(n,1);
Sigma3=zeros(n,1);

P=zeros(n,1);
 %calculating steps
 
 s=(((D1+C1)/2)-(A1-1))/n;
UCS=zeros(n,1)+PolyDa(1,16);

Rgf=[cosd(trend)*cosd(plunge) cosd(plunge)*sind(trend) -sind(plunge); -cosd(rake)*sind(trend)+cosd(trend)*sind(plunge)*sind(rake) cosd(trend)*cosd(rake)+sind(plunge)*sind(rake)*sind(trend) cosd(plunge)*sind(rake); cosd(trend)*cosd(rake)*sind(plunge)+sind(trend)*sind(rake) cosd(rake)*sind(trend)*sind(plunge)-cosd(trend)*sind(rake) cosd(plunge)*cosd(rake)];
Rbf=[cosd(incl)*cosd(azi), sind(azi)*cosd(incl), -sind(incl); -sind(azi), cosd(azi), 0; sind(incl)*cosd(azi), sind(incl)*sind(azi), cosd(incl)];
DP=Pm-Biot*pp;

for i=2:n+1 
   
    Sx(1)=A1-1;
    Sx(i)=Sx(i-1)+s;
    
    
    S3(i-1)=Sx(i-1);
    
    Sf=[xi(i-1) 0 0;0 S2 0;0 0 S3(i-1)];

   %Evaluation of Rg



Sgf=Rgf'*Sf*Rgf;

%Evaluation of Sb

Sbf=Rbf*(Sgf*Rbf');

thetamax(i-1)= double(thetamaxx([azi,incl],[trend,plunge,rake],nu,pp,Pm,xi(i-1),S2,S3(i-1),wbo,FA,Biot));

thetab(i-1)=double((thetamax(i-1)-wbo/2)*3.14/180);

sZZf=Sbf(3,3)-2*nu*(Sbf(1,1)-Sbf(2,2))*cos(2*thetab)-4*nu*Sbf(1,2)*sin(2*thetab)-Biot*pp;
sTTf=Sbf(1,1)+Sbf(2,2)-2*(Sbf(1,1)-Sbf(2,2))*cos(2*thetab)-4*Sbf(1,2)*sin(2*thetab)-Pm-Biot*pp;
sTZf=2*(Sbf(2,3)*cos(thetab)-Sbf(1,3)*sin(thetab));



%Evaluate min tangential stress

st_min(i-1)=0.5*(sZZf(i-1)+sTTf(i-1)-sqrt((sZZf(i-1)-sTTf(i-1)).^2+4*sTZf(i-1).^2));

%Evaluate max tangential stress

st_max(i-1)=0.5*(sZZf(i-1)+sTTf(i-1)+sqrt((sZZf(i-1)-sTTf(i-1)).^2+4*sTZf(i-1).^2));


%Principal stresses

 Sigma1(i-1)=max([st_max(i-1),st_min(i-1),DP]);
 Sigma3(i-1)=min([st_max(i-1),st_min(i-1),DP]);


P(i-1)= Sigma1(i-1)-(UCS(i-1)+Sigma3(i-1)*(tand(45+FA/2))^2);
  
 
end

[Adam]

If n is something valid then the for loop should terminate and return to the calling function anyway. You can easily check what n is by debugging.

[leydy Garcia]

Thanks Adam for your comment. Could you please tell me what you mean by valid? what condition does n have to have to be valid? when I run the code, n is always a rounded number.

Answer 1

[Jan]

Loading data inside by load() consumes a lot of time. Doing this repeatedly is a bad idea and slows down the processing massively. Better use a persistent variable and load the data once.

The next problem is loading the contents if the file directly to the workspace. This is critical if the MAT file contains a variable which is used inside the function also. The dynamic creation of varibales in the workspace impedes the JIT-acceleration, which slows down the processing again.

It is unlikely, that the loop works differently depending from where the function is called. So I assume it is only slow. You can check this easily using the debugger: https://www.mathworks.com/help/matlab/matlab_prog/debugging-process-and-features.html . Set a breakoint and step through the code. As soon as the for loop is ready, the subfunction returns to the caller automatically. You do not have to trigger this explicitly.

[leydy Garcia]

Thanks Jan, but I had already used the debugger and I do not see any error messages. The For Loop finishes and automatically starts again and the subfunction does not return to the caller. When I do not load the data, and use a persistent data all goes fine. The thing is that I have to modify the cripts every time I have to run them using the persistent data. So, if I could only create a new data file every time I need to run the scripts, it would be great.

[Jan]

I'm convinced, that the for loop is not restarted automatically without an explict instruction to do so. Matlab is a deterministic programming language which does not show a random behavior.

I do not understand the meaning of this: "The thing is that I have to modify the cripts every time I have to run them using the persistent data. So, if I could only create a new data file every time I need to run the scripts, it would be great."

[leydy Garcia]

What other way do I have to load the data input? It tried to introduce the input data when running the main function like this:

MySHmaxMatrixPolygon(trend,plunge,rake,nu,pp,Pm,S2,wbo,FA,Biot,azi,incl,Co,x1)

But there is an error that says: "Unbalanced or unexpected parenthesis or bracket." in this line:

[x,fval]=fsolve(@polygon(trend,plunge,rake,nu,pp,Pm,S2,wbo,FA,Biot,azi,incl,Co,x1),x1);

[leydy Garcia]

I managed to avoid loading data inside by load() but still the For loop keeps overruning. I get no error messages because it runs for good. Any help? I have shown the code above but it seems no body sees an error. Thanks

[Jan]

Catch the output of load():

data = load('PolyDa.txt');

Now instead of having the data directly in the workspace, they are available as fields of the struct data.

It is hard to find an error, because you do not mention, what the error is. It is impossible that "the for loops keeps overrunning". Matlab does not behave magically. The loop does exactly what it is instructed to do. If all values from 2 to n+1 are processed, the loop is left and Matlab returns to the caller. As far as I understand your explanations, you claim that the loop is not left regularily. Because this is completely impossible in Matla, you should use the debugger to see, what happens instead.

"A function where I have a FOR loop runs indefinitely when that function is a sub function of the main function." No, this cannot be true. So please examine again, what you consider as problem. Why do you assume, that this loop runs infinitely? Chekc this again, because this assumption cannot be true.

Unfortunately the code is very hard to read in the forum's interface. Limit the line length to improve the readability. I cannot run it, because thetamaxx and thetamax are undefined. What is the input x1? There are many warning message in the editor. What about fixing them? A code with so many warning is not clean.

[leydy Garcia]

THanks Jan. Actually I found no error. I wrote all again whitout leaving long spaces between sentences and very organized. I did not change the code, just wrote all very organized within the FOR LOOP and worked.