check a condition once and remember it later in the code

I want to check for an input at the beginning of my code.

then when i enter a while loop i don't want to recheck this input,

i want to already know it to save time. for example

function [ W,H,T] = AS_R(Q, X,W,H,options)
    % first check
    if strcmp(options.type, 'RSI')
         [ L,R,X_R,X_L ] = RSI(X,r);
    elseif strcmp(options.type, 'GC')
        [ L,R,X_R,X_L ] = GC(X,r,nu);
    else
        disp('Compression type does not exit');
    end
    while (toc<= options.Tmax)
      for i=1:20  
          %do something
      end
        % Do Something
        %--------MY QUESTION ------------------___%
        % At this point i obtain new values, and so i need to call my function
        %  that i checked for in the beginning but dont want to do the strcmp checks again 
        %  I want matlab to already know my choice from the first instance
        
        [ L,R ,X_R,X_L ] = RPI(X,r); % eg it should already use the RSI since it picked it from start.
        k=k+1; 
    end
end

Answer 1

[Adam]

Potentially you can change your if statement to the following, but it depends on what nu is as this does not seem to be defined so would crash anyway on that first call. If it is constant and known in advance then this should work, if it isn't then it is more tricky as you have to pass different arguments to your functions depending which option it is.

if strcmp(options.type, 'RSI')
    func = @RSI;
elseif strcmp(options.type, 'GC')
    func = @(X,r) GC( X, r, nu )
else
    disp('Compression type does not exit');
end

This would allow you to just call

[ L,R,X_R,X_L ] = func( X, r );

later on, if only X and r change in the while loop.

If nu also changes it would need a bit more complicated solution.

[fadams18]

yea would have been perfect but nu changes xD. thanks for this input tho.

[fadams18]

Quick qustion why do you make different assignments

 func = @RSI; 
 %and  
 func = @(X,r) GC( X, r, nu ) 

whats the difference

[Adam]

I could have done the first one in the same way:

func = @(X,r) RSI( X, r )

but it would be redundant.

The difference is in the fact that for my answer nu is a fixed argument known in advance, so it is effectively 'baked into' the function handle when you create it, while X and r are the variable arguments that get passed in later and are not known at function handle creation time.

In the first case there are only variable arguments so the short-form of the function handle works.

func = @GC( nu )

is not valid syntax though and

func = @() GC( nu )

is valid syntax, but not for what you want as it will create a function of 0 arguments, hence if you need to pass in a static argument you need the full thing.

There is a cell array format too, but I have never bothered to learn that as this method has always worked fine for me and I hate cell arrays!

[fadams18]

right thanks!

Answer 2

[fadams18]

So i modified @Adam's answer. I edited the two functions GC and RSI

so that they both have same number of inputs

if strcmp(options.type, 'RSI')
    func = @RSI;
elseif strcmp(options.type, 'GC')
    func = @GC;
else
    disp('Compression type does not exit');
end
[ L,R,X_R,X_L ] = func( X,options);

in that case when if i want GC i pass in eg.

options.type='GC'
options.nu=nu;
options.r=r;

[Adam]

Yep, that works fine for dealing with the issue of different numbers of input arguments, assuming options already contains nu and doesn't have to assign it to the strcut only in the one case.