Error using floor() and ceil() for specific values between o and 1

Good afternoon,

I'm implementing a function that receive a value and look for this value inside of a matrix. The matrix have just integer values, but the value inside the function is not integer. To solve this problem i'm using the following idea:

%The function receive the value of beta;
%find the first index of the matrix that represents a bigger value 
% (Ex: if beta is 8.5, try to find the index that gives the value 9)
index_beta_i_sup = find(Re_36e4 == ceil(beta)); %matrix
beta_i_sup = ceil(beta);
if isempty(index_beta_i_sup) %if the value of beta is not found, I increment +1 until find it
  while isempty(index_beta_i_sup)
    beta_i_sup = beta_i_sup+1;
    index_beta_i_sup = find(Re_36e4 == beta_i_sup);
  end
end

The same idea to try to find the lower index value inside the matrix:

index_beta_i_inf = find(Re_36e4 == floor(beta));
       beta_i_inf = floor(beta);
         if isempty(index_beta_i_inf)
             while isempty(index_beta_i_inf)
                  beta_i_inf = beta_i_inf-1;
                  index_beta_i_inf = find(Re_36e4 == beta_i_inf);
             end
         end

When I run the problem using values: 1 < beta < 180 the function works perfectly.I put some prints to check if I run the code with beta <1, the following error appear.

beta= 8.700000e-01; 
beta_sup= 1;
index_beta_i_sup = 53 %(inside the matrix - OK)

beta= 8.700000e-01; 
beta_inf= 0;
index_beta_i_inf = 52  %(inside the matrix - OK)

| _*%this part is not suppose to be here in the code!*_|
*beta= 104; 
beta_inf= 155;
index_beta_i_inf = 206*  
Error: Index exceeds matrix dimensions.

My matrix has no 206 elements, that's why I have problem, but i didn't intend to make this last calculation, appear alone *(this beta=104). The problem is just for floor().

If I use the value of beta=-0.87, the problem will be in the ceil() function, that will return this beta=104.

Thanks for your support.

[dpb]

Are you simply looking for the nearest integer value in the array to a non-integer input? How about

y = interp1(Y,Y,beta,'nearest');

where Y is your array of integer values, beta is the looked for nearest value to be found.

[Mateus Roberto Urio]

In this case I'm just trying to figure out the index of the nearest value inside the matrix (superior and inferior).But as I mentioned, it works well for values higher than 1, and lower than -1.

I will try to create a new function using interpolation options.

thanks!

Answer 1

[Guillaume]

I'm not too clear on what the problem exactly is, but, if you're trying to find the nearest number above or below your input in the matrix, using a loop over the integers is ... very wasteful.

How about:

%inputs: Re_36e4, a matrix (what an odd and meaningless name for a variable!)
%beta: a number.

%find nearest number above beta
tempRe = Re; %copy Re_36e4 if you don't want to modify it
tempRe(tempRe < beta) = Inf; %replace any  number below beta by +Inf
[beta_i_sup, index_beta_i_sup] = min(tempRe(:));
%note that if beta_i_sup is inf, then there was no number above beta

%find nearest number below beta
tempRe = Re; %copy Re_36e4 if you don't want to modify it
tempRe(tempRe > beta) = -Inf; %replace any  number above beta by -Inf
[beta_i_inf, index_beta_i_inf] = max(tempRe(:));
%note that if beta_i_inf is -inf, then there was no number below beta

The matrix does not even have to be integer for the above to work.

[Mateus Roberto Urio]

Up to now, I just wrote a code to solve a specific problem, but since I'm not an expert with the software, this will be my next step, implement something more efficient . Thanks for your suggestion!

Answer 2

[dpb]

" just trying to figure out the index of the nearest value inside the matrix (superior and inferior)."

Then look for that directly...

I assume Y is ordered --

iLess=find(Y<beta,1,'last');  % last Y < beta location
iPlus=find(Y>beta,1,'first'); % first Y > beta location

If there's no value satisfying the inequality, will return empty, []

[Guillaume]

I believe the OP wants the index of the nearest value greater/smaller than beta, not the first value that happens to be above/below beta.

[dpb]

That's what he'll get with the single caveat of the array being ordered--it's the last under and the first after. Unless they're not ordered those will also be closest but in that case you've got to do a MIN() on the difference to find the nearest.

[Mateus Roberto Urio]

thanks for the help guys! I change a little bit the thinhgs and now its working!