Problem with plotting X @(theta)

Hello there,

I have a function X(theta) defined by two others f(theta) and g(theta). the problem that I have is within plotting the X function. I use fplot command, but it shows me the error: "error: invalid conversion from string to real N-D array error". what does mean this? and how can I properly plot the function I want.

Thank you!

N.B: I use Octave

Answer 1

[Star Strider]

Since you did not share your code, I can only guess what the problem is.

First, if you refer to a function within another function, you must call it as a function, just as you would in any other context.

Second, I am not familiar with Octave and its error messages.

Try this:

f = @(theta) sin(theta);                            % Create Function
g = @(theta) cos(theta);                            % Create Function
X = @(theta) f(theta) .* g(theta);                  % Create Function

figure
fplot(X)
grid

[insaf]

yes, the problem is exactly what you guessed:

hereafter is the code :

I tried the plot and fplot commands but in vain. I tried also using plot(vectorize(inline(x))) (as some friends told me) but no result

theta = linspace(-pi/2,pi/2,0.2);
r0=input ("Enter r0")		
e=input ("Enter epsilon")	
N = input ("Enter desired order")	
k1 = [0:N]
k =find(mod(k1,2)==0)

f = @(theta) (4*r0/pi)*sum(arrayfun(sin(theta*k)/k,[0:N]))+e; 
g = @(theta) (4*r0/pi)*sum(arrayfun(cos(theta*k),[0:N]));
	
x = @(theta) -f(theta)./(g(theta)+1)

[Star Strider]

I believe using arrayfun is inappropriate here. It seems that what you want is to simply do vector multiplication to produce a matrix, then sum that.

Try these:

r0=input ("Enter r0")		
e=input ("Enter epsilon")	
N = input ("Enter desired order")	
k1 = [0:N];
k =find(mod(k1,2)==0)
f = @(theta) (4*r0/pi)*sum(sin(theta*k(:))./k(:))+e;
g = @(theta) (4*r0/pi)*sum(cos(theta.*k(:)));
	
x = @(theta) -f(theta)./(g(theta)+1)

figure
fplot(x, [0 10])
grid

It still throws a warning (in MATLAB). It nevertheless produces the plot, and that is the desired result.

Note that the ‘(:)’ subscript notation forces a column vector. If Octave does not have that option, use a simple transpose on ‘k’ instead.

This assumes that ‘theta’ as supplied to the function will always be a row vector.

[insaf]

I used a transpose on ‘k’ as you told me, but here is another error "error: operator *: nonconformant arguments (op1 is 5x1, op2 is 5x1)"

for sum(arrayfun(f)), I use it to sum all over the values of k, I don't know if using simply sum will do the same thing

[Star Strider]

Using sum in this context is likely appropriate.

The transpose may not be necessary, since fplot may have internal loops, and do element-wise operations by default.

This works for me (although MATLAB still throws warnings), and produces the plot:

f = @(theta) (4*r0/pi)*sum(sin(theta*k)/k)+e;
g = @(theta) (4*r0/pi)*sum(cos(theta*k));

The rest of the code in my previous Comment is unchanged.

EDIT —

The fplot function provides values for ‘theta’ here. The range is defined by the values in the vector [0 10].

Note that your linspace call produces an empty vector, at least in MATLAB.

The Plot —

[insaf]

in octave I don't get a plot, and the error still persists,

error: operator *: nonconformant arguments (op1 is 5x1, op2 is 6x1)

error: called from

starcomment>@<anonymous> at line 6 column 50

starcomment>@<anonymous> at line 9 column 23

fplot at line 136 column 8

starcomment at line 11 column 1

[Star Strider]

I will help you with this as much as I can.

Assuming that fplot considers ‘theta’ to be a row vector, transposing it to a column vector and keeping ‘k’ as a row vector to do the multiplication may work:

f = @(theta) (4*r0/pi)*sum(sin(theta'*k)/k)+e;
g = @(theta) (4*r0/pi)*sum(cos(theta'*k));

I looked at the Octave fplot help documentation. It gave no examples of something like this, and did not say how it worked internally.

[madhan ravi]

+1 this answer works for me without error

[insaf]

Star Strider and madhan ravi thank you for your patience and help, I'll re-test the functions you give me and I look carefully at the size of the matrices while multiplying, maybe the error comes from this point

Thank you again

[Star Strider]

Our pleasure.

It would be nice to know how Octave’s fplot does its calculations. It appears not to to be as robust to row and column orientations as the MATLAB fplot function is.

[madhan ravi]

Anytime :) , second Star Striders point.

Answer 2

[madhan ravi]

str2double() % to convert string to double and then plot

[insaf]

Thank you for replying

I tried your suggestion, but this is not working, it shows me this "error: unary operator '.'' not implemented for 'function handle' operands"

[madhan ravi]

please upload your code and datas

[insaf]

here is the code

theta = linspace(-pi/2,pi/2,0.2);
r0=input ("Enter r0")		
e=input ("Enter epsilon")	
N = input ("Enter desired order")	
k1 = [0:N]
k =find(mod(k1,2)==0)
f = @(theta) (4*r0/pi)*sum(arrayfun(sin(theta*k)/k,[0:N]))+e; 
g = @(theta) (4*r0/pi)*sum(arrayfun(cos(theta*k),[0:N]));
	
x = @(theta) -f(theta)./(g(theta)+1)

[madhan ravi]

theta = -pi/2:.2:pi/2;
r0=3;   
e=5 ;
N = 7;  
k1 = 0:N;
k =find(mod(k1,2)==0);
f =  (4*r0/pi)*sum(sin(theta.'*k)/k)+e; 
g = (4*r0/pi)*sum(cos(theta.'*k));   
x = -f./(g+1)
plot(x)

[insaf]

that's work! but for theta, is there any difference between the way I defined it (with linespace) and the way you write it ?

[madhan ravi]

0.2 is the step , in matlab 0.2 (should be an integer ) so I assumed you wanted the step 0.2 , the place where you put 0.2 in linspace represents number of intervals from the starting point to the end point.

• see linspace in matlab

[KALYAN ACHARJYA]

@insaf, Reagrding theta statement: both are same meaning, just representations are different.

Read Here

[madhan ravi]

Thanks Kalyan

See the illustration below in MATLAB (no idea about octave's linspace though):

>> linspace(-pi/2,pi/2,0.2)

ans =

  1×0 empty double row vector

>> -pi/2:0.2:pi/2

ans =

  Columns 1 through 7

   -1.5708   -1.3708   -1.1708   -0.9708   -0.7708   -0.5708   -0.3708

  Columns 8 through 14

   -0.1708    0.0292    0.2292    0.4292    0.6292    0.8292    1.0292

  Columns 15 through 16

    1.2292    1.4292

>> 

[insaf]

thank you @madhan ravi and @KALYAN ACHARJYA for your help! I appreciate that.

If I could ask, how can I add a loop "for" so it asks me each time to enter a new value for N, calculate and represent the new figure on the last one ?

[madhan ravi]

The simplest way is to convert it as a function and just give the input N like shown below:

for i = 1:10
N=input('what value  ? for N: ','s'); % if you just enter without anything the loop will stop
if isempty(N)==1
    break
else
main(N) % function call
end
end

function main(N)    % function definition
theta = -pi/2:.2:pi/2;
r0=3;   
e=5 ;
k1 = 0:str2double(N);
k =find(mod(k1,2)==0);
f =  (4*r0/pi)*sum(sin(theta.'*k)/k)+e; 
g = (4*r0/pi)*sum(cos(theta.'*k));   
x = -f./(g+1);
figure
plot(x)
end

[Star Strider]

@insaf —

Please note that I figured out the reason your code was not working, and got it to work in my Answer.

[Image Analyst]

Could it be that your solutions are using MATLAB and he's not? He's using Octave. I'm not sure exactly how compatible it is, like do function names match up exactly, etc.?

[Star Strider]

My solution apparently works in Octave.

[madhan ravi]

@sir Image Analyst agree sometimes the OP is lucky but mostly not , functions are different as you mentioned.

[insaf]

thank you Star Strider for your response, effectively, I noticed the error in defining theta. for the other code using "For", it doesn't work in octave since the "main" function is not defined there. I will search for an equivalent for it in octave, I think it will be easy then to implement a loop "for".

Mr madhan ravi what I'm wondering is the values I got using your code for f and g: as theta is defined in a large interval, f should then be a function of theta, but the code gives me a number! I guess because it calculates the summation along theta and k, while I want it to be just on k, so the f will be a function of theta.

the same thing with g, and x of course.

I hope my problem is clear for you

[madhan ravi]

so see Star Striders answer