False Output from find-function?

Heyho

I'm looking for maximum values of a table and am using

[rows, columns] = find(A==max(A));

to do so in the attached matrix.

It gets me this:

     1     1
     2     2
     3     3
     4     4
     5     5
     6     6
     7     7
     8     8
     9     9
    10    10
    11    11
    12    12
    13    13
    14    14
    15    15
    16    16
    19    17
    17    18
    18    18
    19    19
    20    20

The matrix is 20x20 and the output has 21 values. :S

Also in my opinion the line...

    19    17

...isn't true, cause I can't see a maximum there? This shouldn't be there at first place, should it?

Before and after this everything looks fine to me.

Why is this and how can I prevent that from happening?

Answer 1

[Thiago Henrique Gomes Lobato]

max(A) is a vector, not a number:

max(A)
ans =

  Columns 1 through 15

    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    1.0000    0.9999    1.0000    1.0000    1.0000    1.0000

  Columns 16 through 20

    1.0000    0.7950    1.0000    1.0000    0.9998

The value at A(19,17) is the 0.795, which is not a maximum of the matrix but rather a maximum of the column.

if you want to get the maximum of the whole matrix you need this:

max(A(:))
ans =

    1.0000

And when you compare, remember that float value comparison is everything but reliable, so you need to have a given tolerance.

In your code it would be something like this:

[rows, columns] = find( abs(A-max(A(:)))<1e-4  );
ans =

     1     1
     2     2
     3     3
     4     4
     5     5
     6     6
     7     7
     8     8
     9     9
    10    10
    11    11
    12    12
    13    13
    14    14
    15    15
    16    16
    17    18
    18    18
    19    19

[Rene]

Hey Thiago

I want to put the machting "rows and numbers" to a table, so just a single value isn't what I'm looking for.

The last version with the tolerance is the way I need to go. But you're right, I forgot that all these "1" aren't exactly 1.

But why is the last one missing in the code with the tolerance? (the 20 20 is a maximum too)

Edit: Got it. It was cause of the tolerance value.

[Rene]

I'm sorry but it did not work out like I wanted it.

I tried it with another matrix:

A = [1 0.5 0.7;0.4 0.8 0.4;0.2 0.3 0.7];

It only gave me the value 1.

I then created a loop to get the maximum of each row

k = 1;
for i = 1 : size(A,1)
    Max(k,1) = max(A(i,:));
    k = k+1;
end

This gives me the maximums I need in the right order.

1
0.8
0.7

Now I want the location too. Like so:

%row   col     value
1	1	1
2	2	0.8
3	3	0.7

I tried the find again but as soon as there are multiple of the same value in the matrix it wont work like this:

k = 1;
for i = 1 : size(A,1)
    Max(k,1) = max(A(i,:));
    [row,col] = find(A==max(A(i,:)));
    row(k,1) = row;
    col(k,1) = col;
    k = k+1;
end

Any idea how to get the column value of the exact max value I'm looking for?

[Thiago Henrique Gomes Lobato]

This will probably work for you

A = [1 0.5 0.7;0.4 0.8 0.4;0.2 0.3 0.7];
Amax = max(A); % Get maximum from each column
DiffFromMax  = A-Amax; % Zero values are maximum locations
tol = -1e-4; % Note that the tolerance is negative
[rows, columns] = find( DiffFromMax>tol )