Is there command to go line I want?."goto"

I usually work with Fortran, C++ and there is command "goto" which goes to line I need. Now in my matlab 12 I could not do it.please help me, thanks for response.

[dpb]

Besides the previous comment on how to refactor, for Matlab you've got a lot of stuff like the following--

E1=23.*VR*13.605;
E2=4.1*VR*13.605;
E3=1.3*VR*13.605
...
S1=F1*VA*(LOG(S0/E1)-BC2)/E1;
S2=F2*VA*(LOG(S0/E2)-BC2)/E2;
S3=F3*VA*(LOG(S0/E3)-BC2)/E3;

that should be rewritten to use the array notation more like

E=13.605*[23.0 4.1 1.3].'*VR;
...
S=F.*VA.*(LOG(S0./E)-BC2)./E;

etc., ...

Then address the elements of E and S as E(1), S(1), where/if needed.

Also, you're aliasing the builtin Matlab function pi with a less accurate constant (which isn't apparently used, anyway). Don't do this kind of thing in Matlab; it'll bite nastily at some point if you get in the habit.

Answer 1

[Azzi Abdelmalek]

There is no such command in Matlab, plus, it's not recommended to use goto. It's hard to read a big code with many goto inside.

You can replace goto by using a better structured code, with command like: while, calling functions, ...

[Vachik]

what kind of functions? IN my code it is necessary, here is my code , if you can help me to modify this for matlab please make comment.

rmpi=139.567;    
rmp=938.27;
rmhe4=3727.381;
rmhe3=2808.392;
rmh3=2808.923;
rmh2=1875.614;
rm=rmhe4;
rm2=rm*rm;
rp11=rmp;
re12k=6.217;
re122=rm+re12k;
ZFR=2;
VR0=1.0;
DT=0.000010;
R0=VR0;
RL0=DT;
RT1=VR0*DT;
VR=1.406;
RME=0.511;
E1=23.*VR*13.605;
E2=4.1*VR*13.605;
E3=1.3*VR*13.605;
F1=2/6;
F2=2/6;
F3=2/6;
VA5=0.0768;
A=VA5;
B=18;
VA1=VA5;
for i=1:10;
re12=re122;
RT=RT1;
pi=3.141592654;
r=rand;
de=0.0;
tp=0.0;
VRLL=0.0;
%10=39
g12=re12/rm;
g122=g12*g12;
b122=1.-1./g122;
b12=sqrt(b122);
BC2=b122;
GL2=g122;
A2=VA1/BC2;
VA6=ZFR*ZFR*A2;
rp12=re12*re12-rm2;
WMX=2.*RME*BC2*GL2;
WMX=WMX*1.E6;
VA=VA6*1.E6;
S0=WMX;
if S0<=E1;
F1=0;
end
if S0<=E2; 
F2=0;
end
if S0<=E3;
end
F3=0;
S1=F1*VA*(LOG(S0/E1)-BC2)/E1;
S2=F2*VA*(LOG(S0/E2)-BC2)/E2;
S3=F3*VA*(LOG(S0/E3)-BC2)/E3;
S=S1+S2+S3;
R1=S1/S;
R2=(S1+S2)/S;
R3=(S1+S2+S3)/S;
if WMX<=E1;
F1=0;
end
if WMX<=E2;
F2=0;
end
if WMX<=E3;
F3=0;
end
C1=F1*VA*(1./E1-1./WMX-BC2*LOG(WMX/E1)/WMX);
C2=F2*VA*(1./E2-1./WMX-BC2*LOG(WMX/E2)/WMX);
C3=F3*VA*(1./E3-1./WMX-BC2*LOG(WMX/E3)/WMX);
C=C1+C2+C3;
RN1=C1/C;
RN2=(C1+C2)/C;
RN3=(C1+C2+C3)/C;
SS=C+S;
RG1=S/SS;  
RR=rand(1);
DD=LOG(RR)/(-SS);
VRLL=VRLL+DD;
if VRLL>=RT; 
 GO TO 50
end
RN=rand(1);
if RN<=RG1; 
 GO TO 20
end
RN=rand(1);
if RN>=RN2;
 GO TO 13
end
if RN>=RN1; 
 GO TO 14
end
EE=E1;
GO TO 15
%13=106  
GO TO 15
%14=108  
%15=109  
EE1=EE;
%16=111  
EE=EE1/(1.0-RN+RN*EE1/WMX);
RN=rand(1);
if RN>=(1.-BC2*EE/WMX); 
GO TO 16
end
GO TO 30
%20=118  
if RN>=R2; 
GO TO 25
end
if RN>=R1;
GO TO 26
end
EE=E1;
GO TO 30
%25=127  
GO TO 30
%26=129  
%30=130  
de=de+EE*1.E-6;
re12=re12-EE*1.E-6;
if re12-rm<=0.05;
GO TO 50 
end
tp=tp+DD/b12;
GO TO 10
%50=138  
re22k=de;
rp22=re12;  
r1=randn(1);
r2=randn(1);
rp222=rp22*(1.+0.0003*r1);
re22k=re22k+0
rp22=rp22+0

%11=146

end

[dpb]

...In my code it [GOTO] is necessary...

It is always possible to restructure/refactor code to eliminate GOTO. Depending on the actual logic, it may be as simple as IF...ELSE...END or a WHILE...END; in other cases (particularly often error handling) it may need more serious refactoring of the code. try...catch is often useful for such cases.

As for your particular code, two suggestions --

1) Keep the Fortran version but either run the standalone via system or turn it into a mex function, or

2) If it is mandatory to convert to Matlab native code then at least provide a clean snipped version that contains only the relevant sections you're having difficulty with in the native Fortran including line numbers. It's impossible in what you've give to determine where the branches are going and there are large chunks that don't appear to have any bearing on the question.

ADDENDUM:

OK, as an aid to get you started, I snipped the code down to just the branches and a line or two between so it would fit in a page to see the overall structure. Then I did a couple of the transformations as starters...

        for i=1:10;
        ...
        VRLL=0.0;

%   10   continue
      flag10=true;
      while flag10
        DD=LOG(RR)/(-SS);
        VRLL=VRLL+DD;
        %if VRLL>=RT; GO TO 50, end
        if VRLL>=RT, continue, end  % skip to next iteration
        RN=rand(1);
        %if RN<=RG1; GO TO 20, end
        if RN>RG1  % execute next instead skipping on negated logic
          RN=rand(1);
          if RN>=RN2; GO TO 13, end
          if RN>=RN1; GO TO 14, end
          EE=E1;
          GO TO 15
 13       continue
          GO TO 15
 14       continue
 15       continue
          EE1=EE;
 16       continue
          EE=EE1/(1.0-RN+RN*EE1/WMX);
          RN=rand(1);
          if RN>=(1.-BC2*EE/WMX); GO TO 16, end
          GO TO 30
        end  % 20  target end ←
        if RN>=R2; GO TO 25, end
        if RN>=R1; GO TO 26, end
        EE=E1;
        GO TO 30
 25     continue
        GO TO 30
 26     continue
 30     continue
        de=de+EE*1.E-6;
        re12=re12-EE*1.E-6;
        if re12-rm<=0.05; GO TO 50, end
        tp=tp+DD/b12;
      end  % GO TO 10 conditional loop
 50 continue
      ...
    end

1) the GOTO 10 is just an indefinite loop -- I replaced it with a WHILE...END clause where I defined a flag variable that may or may not in the end be of any need; I didn't go far enough to decipher just how the loop does finally get terminated so where setting the flag variable false is the right spot, specifically.

2) Then, there's no GOTO 11 so that label can be eliminated which leads to...

3) the GOTO 50 is just causing a skip of the body of the for...end loop which is, in Matlab continue so the line

if VRLL>=RT, GOTO 50, end

becomes

if VRLL>=RT, continue, end

There's another of these later on; when that one's taken care of the same way then the label 50 can be dispensed with.

4) The branch

if RN<=RG1; GO TO 20, end

causes conditional skipping of the code to label 20 which is the same thing as the conditional execution of the code as

    if RN~<=RG1; {CODE BLOCK}, end

I just changed the conditional from <= to > and added the end at the target label and indented for legibility...

if RN>RG1  % execute next instead skipping on negated logic

The remainder is just a continuation of the exercise as above.

[John D'Errico]

I totally agree with dpb. You can always formulate your code NOT to use goto.

If you think you need it, sorry, but you are just avoiding the task of learning the tools present in MATLAB.

Answer 2

[Vishwanath Salokye]

Hi,

Their is no such command in MATLAB. using function call you can jump to the particular location. But in matlab their are two types of workspace one is base work space and other one is function workspace. When you enter in to any function base workspace will be empty you have to use "evalin" command to get base workspace value to function workspace

[Vachik]

how to use evalin?

[dpb]

For this purpose don't even think about it... :)

Answer 3

[Guruprasad S]

Hi, I wrote a function to read the input whole number between 2 and 6. The conditions are, if the input number is <=2 or >=6, it must repeat the function with an error message. Similarly, if the number entered is a fraction, it must repeat the function with an error message. This might be an answer to goto function.

function new_file
x = input('input a whole number between 2 and 6  ');
 if x==3
    fprintf('three');
elseif x==4
    fprintf('four');
elseif x==5
    fprintf('five');
 elseif x<=2 || x>=6;
    fprintf('read the input properly\n');
   new_file
 else
     fprintf('input whole number only\n')
     new_file
 end

[Steven Lord]

There is no goto function in MATLAB.

For this application rather than having your function call itself recursively just use a while loop. while the user has not entered a valid input, ask them for a valid input.

[dpb]

Besides @Steven Lord's suggestion on refactoring/restructuring, note that the above function never returns the input value -- while perhaps not the best use of recursion, with that correction looks like it should work just at a glance...excepting you'll then also return all the wrong answers besides the last one unless you do something about that.

Also, use disp() instead of fprintf for your debugging of the results -- your lines have no \n so will run on to the command line prompt.

While I still would no recommend it, the following solves the problems --

function x=new_file
  x = input('input a whole number between 2 and 6  ');
  if x==3
    disp('three');
  elseif x==4
    dips('four');
  elseif x==5
    disp('five');
  elseif x<=2 || x>=6;
    fprintf('read the input properly\n');
    clear x
    new_file
  else
    fprintf('input whole number only\n')
    clear x
    new_file
  end
end