How to define a variable?

Dear all,

I have this code:

    prompt = {'Type the value j:'};
    dlg_title = 'Create your own model';
    num_lines = 1;
    j = 0.5;
    def = {'0.5'};
    answer = inputdlg(prompt,dlg_title,num_lines,def);
    imdl =  mk_common_model('d2d1c',16);
    img_1 = mk_image(imdl);
    figure
    show_fem(img_1);
    img_2 = img_1;
    ['x-' num2str(j) '.^2+ (y- 0.5) .^2<0.1^2, ''x,'  'y,'  'z']
    select_fcn = inline('(x-j).^2+(y-0.5).^2<0.1^2','x','y','z')
    img_2.elem_data = 1 + elem_select(img_2.fwd_model, select_fcn);
    figure
    show_fem(img_2);
    vh = fwd_solve(img_1);
    vi = fwd_solve(img_2);

I have a problem with variable j. I want to change my inline function, for example user enters j = 0.1 so my inline function should look like (x-0.1).^2+(y-0.5).^2<0.1^2','x','y','z'. But Command Window returns this:

ans =
x-0.5.^2+ (y- 0.5) .^2<0.1^2, 'x,y,z

select_fcn =

       Inline function:
       select_fcn(x,y,z) = (x-j).^2+(y-0.5).^2<0.1^2

Does anyone know where is the mistake?

Thank you for your ideas.

[per isakson]

Doc says:

    inline will be removed in a future release. 
    Use Anonymous Functions instead.

• Any special reason not to use Anonymous Functions?
• Which Matlab release do you use?

[Stephen23]

Use Anonymous Functions instead of inline.

Also note that you should avoid using j and i as the names of variables, as these are the names of the inbuilt imaginary unit .

[Veronika]

So I rewrote the inline function to an anonymous function. And I rewrote the variable j to the variable p. Now, I have this code:

   prompt = {'Napište hodnotu p:'};
dlg_title = 'Vytvoření vlastního modelu';
num_lines = 1;
p = 0.5;
def = {'0.5'};
answer = inputdlg(prompt,dlg_title,num_lines,def);
imdl =  mk_common_model('d2d1c',16);
img_1 = mk_image(imdl);
figure
show_fem(img_1);
img_2 = img_1;
['x-' num2str(p) '.^2+ (y- 0.5).^2<0.1^2']
select_fcn = @(x,y) (x-p).^2+(y-0.5).^2<0.1^2;
img_2.elem_data = 1 + elem_select(img_2.fwd_model, select_fcn);
figure
show_fem(img_2);
vh = fwd_solve(img_1);
vi = fwd_solve(img_2);

But Comand Window returns this:>> GUI

ans =

x-0.5.^2+ (y- 0.5).^2<0.1^2

Error using C:\Users\Veronika\Desktop\BP\GUI.m>@(x,y)(x-p).^2+(y-0.5).^2<0.1^2
Too many input arguments.

Error in elem_select (line 57)
    memb_frac = mean( feval(select_fcn,x,y,z), 2);

Error in GUI>pushbutton5_Callback (line 115)
img_2.elem_data = 1 + elem_select(img_2.fwd_model, select_fcn);

Error in gui_mainfcn (line 96)
        feval(varargin{:});

Error in GUI (line 42)
    gui_mainfcn(gui_State, varargin{:});

Error in @(hObject,eventdata)GUI('pushbutton5_Callback',hObject,eventdata,guidata(hObject))

Error while evaluating uicontrol Callback

My release of Matlab is MATLAB R2011b.

[Stephen23]

It really depends on the function elem_select, which I could not find any references to online or within my own MATLAB. Did you write this function yourself, or obtain it from someone, or are you using a third-party toolbox?

There are several possibilities:

• perhaps it requires a z input as well as the x and y inputs. You will need to check the source/documentation for this function to check this.
• maybe the two standard callback inputs hObject and eventdata are somehow conflicting with the inputs that you are defining.

Can you tell use something about elem_select ? Or upload it in a new comment?

[Veronika]

Function elem_select comes from EIDORS, specifically from this site: http://eidors3d.sourceforge.net/tutorial/EIDORS_basics/forward_solvers.shtml.

[Veronika]

I rewrote code this:

    prompt = {'Napište hodnotu p:'};
dlg_title = 'Vytvoření vlastního modelu';
num_lines = 1;
p = 0.5;
def = {'0.5'};
answer = inputdlg(prompt,dlg_title,num_lines,def);
imdl =  mk_common_model('d2d1c',16);
img_1 = mk_image(imdl);
figure
show_fem(img_1);
img_2 = img_1;
['x-' num2str(p) '.^2+ (y- 0.5).^2<0.1^2, *''x,'  'y,'  'z'* ]
select_fcn = @(x,y, *z* ) (x-p).^2+(y-0.5).^2<0.1^2;
img_2.elem_data = 1 + elem_select(img_2.fwd_model, select_fcn);
figure
show_fem(img_2);
vh = fwd_solve(img_1);
vi = fwd_solve(img_2);

This really works without errors. But I needed to make me changing variable p whichever user inserts into the prompt.For example user inserts 0.1 so function will be: (x-0.1).^2+ (y- 0.5).^2<0.1^2.

[Stephen23]

And what kind of an object is img_2.fwd_model ?

At this point I think your best bet would be to start using the debugger , and set a breakpoint inside the Anonymous function.

You could also try this anonymous function instead:

select_fcn = @(varargin)disp(nargin);

Just to see how many inputs it is getting. But the debugger is going to be your best tool for solving this.

[Veronika]

img_2.fwd_model is figure of the conductivity based on the function (x-p).^2+(y-0.5).^2<0.1^2.

Ok, I try debugger.

Answer 1

[per isakson]

"[...]changing variable p whichever user inserts" &nbsp I assume "*z*" is a highlighted input variable. However, I cannot find any comment on "*z*".

The anonymous function (function handle) includes a "snapshot" of the current workspace of the moment when it is created. In your case p with the value 0.5 is stored in select_fcn. Later changes of the value of p doesn't doesn't affect select_fcn.

Proposal: replace

    select_fcn = @(x,y, *z* ) (x-p).^2+(y-0.5).^2<0.1^2;

by

    select_fcn = @(x,y,p) (x-p).^2+(y-0.5).^2<0.1^2;

&nbsp

"Error in elem_select (line 57) memb_frac = mean( feval(select_fcn,x,y,z), 2);" &nbsp Comments:

• feval(select_fcn,x,y,z) select_fcn is defined with two input arguments, @(x,y)
• feval is not required to evaluate an anonymous function. Replace feval(select_fcn,x,y,z)) by select_fcn(x,y,z)

[Veronika]

Thank you for your answer.

1. I rewrote z to p, but nothing changed, still returns me to a value of 0.5 rather than a user specified value.
2. And for the second error, I do not know where I have to change that select_fcn(x,y,z), because I can´t see it in my code.

[per isakson]

What is the intended role of z in

    select_fcn = @(x,y,z) (x-p).^2+(y-0.5).^2<0.1^2;

z is input, but not used. Why?

I looked at the EIDORS page, but that didn't help much. I didn't give it the time needed. My conclusions:

• EIDORS is a large toolbox, which is 10++ years old. Anonymous functions didn't exist in Matlab at that time. (However, the current release is only a year old.)
• You try to call the function elem_select with a set of input arguments that doesn't honor the documentation. No surprise an error is thrown.
• You did not provide this context in your question and thus it was hard to give useful answers.

"because I can´t see it in my code" &nbsp That's because elem_select (The name is in the error message.) is a function in the EIDORS toolbox.

&nbsp

It is not a good idea to modify code in the EIDORS toolbox. Thus, it remains to adhere to the documentation of EIDORS. (We may trust The MathWorks that inline can be replaced by anonymous functions.) The definition of the anonymous function must be like

    select_fcn = @(x,y,z) (x-0.5).^2+(y-0.5).^2<0.1^2;

EIDORS doesn't say one may add an extra input argument. That's mean one cannot.

&nbsp

One solution is to redefine the anonymous function for each new value of p.