Why won't my if statement execute in the for loop?

I have the following code which sets variables used to a for loop:

 %Find the entry of P that lambda is most sensitive to, i.e. the largest
 %value in the elasticity matrix. M finds the largest entries in the columns of P and puts them into
 %a row vector. I returns the row index of the elements of M found in E:

 [M,I] = max(E); 
 disp(M)
 disp(I)
 index=find(M==max(M));        %Find index of largest element in M. This corresponds to the column of E this element  was found in
 disp('most sensitive found in column')
 disp(index)
 mostSensitive=E(I(index),index);  %Set mostSensitive to the largest value found in M

and I wish to execute the if statement in the following for loop:

%Now create a vector which counts how many times mostSensitive appears in
%the elasticity matrix:

 numAppearances = 0;
 rowIndices=[];
 for k=1:size(M)   %Traverse vector of maximum values of columns of E
    disp('We are at index')
    disp(k)

    disp('We are at element')
    disp(M(k))
    disp(mostSensitive)

    if M(k)==mostSensitive    %If the current element in the vector of the largest elements of S is equivalent to the highest one found,
        disp('In if statement suite')
        numAppearances=numAppearances+1;
        rowIndices(k)=I(k);
    end
end

When the interpreter reaches the if statement, it will not execute. I have taken the followng steps:

1. MATLAB documentation has not informed me on why this may be happening.
2. I have used error checks using disp(x) and have seen that the if condition is true
3. I have also used the debugger
4. I have looked at similar questions on MATLAB answers and generally googled solutions

Answer 1

[Image Analyst]

If it does actually execute statements in the for loop, but the "if" condition is false, then I suspect you're comparing floating point numbers. And you'll have to read the FAQ: http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

Look at what your disp statements put out into the command window. Does M(k) or mostSensitive have any fractional components? If so, then the FAQ definitely applies. What does this say:

whos M
whos mostSensitive

If either is a double, then you might use tolerances like the FAQ shows you, or else cast them to int32 if you can.

[MATmagician]

Thank you. I thought it was an issue with floating point numbers. I will do some tinkering with what you've shared and see if this fixes it

[MATmagician]

The issue is now resolved. I now know how to compare floats properly. Thank you again

[Image Analyst]

You're welcome. Granted, it's not obvious so don't feel bad. Thanks for Accepting.

Answer 2

[Image Analyst]

My guess is that M is empty so that it never enters the for loop. Why can you not discover that it never steps into the for loop if you're using the debugger? Are you sure you know how to use it?

[MATmagician]

M is not empty. I have it displayed right after it is made. It does enter the for loop. I have error checks which inform me that it does

Answer 3

[Walter Roberson]

You have

   for k=1:size(M)   %Traverse vector of maximum values of columns of E

that needs to use length(M) instead of size(M) . Remember that size(M) is a vector, and you cannot use a vector as a limit in a colon operator.

[MATmagician]

Thanks. I forgot to correct that.