Extracting elements from one matrix according to criteria from another matrix

I have the following problem: I have say 100 3-dimensional points, and a correlation matrix A with the correlation of all the points to each other, hence A has the dimensions 100 x 100.

For all my points I have another matrix B with the coordinates of each point, such that B is a 3 x 100 matrix.

Now I want the following to be done: From my correlation matrix A, I only want to have the correlation of points extracted, where the x coordinate of the point is 0 or bigger than zero, such that I arrive at a new correlation matrix, and also for the matrix B I only want to keep the points which fullfill this criterion. The x values are stored in the first row of the matrix B for each of the 100 points.

How could I do that?

Answer 1

[the cyclist]

If I understand correctly what you want, I think this does it.

index = B(1,:)>=0;

A = A(index,:);
A = A(:,index);

B = B(:,index);

[MiauMiau]

I do not get the correct result with your code.

Given I have B = [ 3 -2 -5 4; 0 9 8 7; 1 2 3 4], I would want to keep the first and fourth column. Now I only get:

>> index = B(:,1)>=0

index =

     1
     1
     1

and hence

>> B = B(index,:)

B =

     3    -2    -5     4
     0     9     8     7
     1     2     3     4

[the cyclist]

Sorry, I thought your B was 100x3. I corrected my code.

[MiauMiau]

Ok, thanks! B is correct now, but A isn't. So for A, since in B only column 1 and 4 are kept, we would need to keep only the correlation between the first point and the fourth. These are encoded in A in (1,1) (correlation of the first point with itself), (4,4) (correlation of the fourth point with itself), then (1,4) and (4,1). How would we do that?