how to identify number of values in a column vector that are not zero values and are not NaN values?

Hello,

Lets say I have a 2458 x 1 column vector in a cell array, A, that contains zeros, nonzeros, and NaN values. I would like to identify how many values are nonzero values and are not NaN values. This means that I have to filter out the zero values and NaN values. This is the current code I'm using but this portion, (A{1,1} ~= NaN), seems to not be working.

logical = (A{1,1} ~= 0) & (A{1,1} ~= NaN);
A{1,1}(~logical) = [];
length(A{1,1})

When I run the code, it has a 1 value at the row indices that contain NaN, meaning that the row index value is true...But it shouldn't be because it is a NaN value and I want to only determine the values that are nonzero and are not NaN values. I was wondering why this does not work and what I can do to fix this? I'm using 2015a btw and any help or advice would be appreciated.

Answer 1

[Star Strider]

I don’t have your cell array to experiment with, but this works for my synthesised one:

A = {1, 3, 5, NaN, 7, 11, 13, 0 17};
logc = @(x) (x ~= 0) & ~isnan(x);
Idx = cellfun(logc, A);

[George Vuong]

Hello again Star Strider, thanks for your quick response.

So once the logical array is outputted, how would you go about determining the number of ones in the logical array? I would do something like this:

 length(find(logc == 1))

or

 A(~logc = []) 
 length(A)

Or do you prefer another method?

[Star Strider]

My pleasure.

To determine the number of ones in the vector, the sum or nnz functions will both work:

NonZeroEmenents = sum(Idx);
NonZeroElements = nnz(Idx);

You have to use the ‘Idx’ vector, not the ‘logc’ function.

[George Vuong]

Never knew about nnz function, which is pretty neat. And summing the vector is another great idea. I'm glad I asked you that question even though I already had other feasible solutions as stated in my first comment. Thanks again!

[Star Strider]

My pleasure!

Thank you!