How to make conditional calculation?

I have two vectors, contained in matrix M, each with a dimension of 1000 by 1. Each vector contains integer numbers. The values in one vector (call it vector two) is generated in way that it depends on the values in the other vector (call it vector one). I want to calculate the variance of the numbers in vector two if the values are below a certain threshold in vector one. I try to use the code below, but it returns nothing. Where am I making the mistake?

if M(:,1) < 0
    var(M(:,2))
end

Answer 1

[Image Analyst]

Maybe you mean this:

indexesToConsider = M(:, 1) < 0; % Consider the index if the M value is less than 0.
col2Variance = var(M(indexesToConsider, 2)); % Compute variance of only those rows.

[Snoopy]

This the logical indexing Walter above mentions right? I now use it. Thanks a lot.

[Image Analyst]

Yes, it's logical indexing in that it gives values of true and false. If you passed that into find() you'd get "linear indexes" which are the actual index numbers, like [1, 3, 5, 33, 45, etc.] instead of [1 0 1 0 1 0 0 etc.]

Answer 2

[Walter Roberson]

if M(:,1) < 0

means the same thing as

if all(M(:,1) < 0)

which is probably not true.

Look up "logical indexing"

[Snoopy]

Thanks. I now use the find function which returns the row numbers (or the row index I shall say?) where the condition for vector one is satisfied. Then I use these row numbers as the row subscript in vector two, and carry out the operation on this subset of the entries of vector two. Is this perhaps the correct approach?

[Walter Roberson]

That is one approach, but it is more efficient to use logical indexing. https://www.mathworks.com/help/matlab/math/matrix-indexing.html#bq7egb6-1