Standard Deviation of 3D matrix with NaN elements

I have a 3D matrix of size (170 x 170 x 27). Some of the entries are NaN, zero and the rest have real values.

I would like to obtain the standard deviation of this matrix, excluding the NaNs. I have tryed several commands, but I haven't been able to get this to work.

Could you help me?

Thank you in advance,

Antonio :)

Answer 1

[Johannes Fischer]

As described in the doumentation for std, there is a flag on how to deal with 'NaN' values. If you want the standard deviation of all values, excluding the NaNs, then write

% dummy matrix
m=rand(170, 170, 27);
m(m>0.9) = NaN;

sdev = std(m, 0, 'all', 'omitnan')

[António Rouxinol Fragoso]

Thank you for your answer! Unfortunatly, it doesn't seem to solve the problem. The following information appears:

Error using size
Dimension argument must be a positive integer scalar within indexing range.

Error in var (line 109)
n = size(x,dim);

Error in std (line 51)
y = sqrt(var(varargin{:}));

Error in untitled2 (line 3)
sdev = std(m, 0, 'all', 'omitnan');

[Rik]

What release are you using? In previous releases the syntax for this function was slightly different.

[António Rouxinol Fragoso]

I'm using R2018a.

[Rik]

The 'all' syntax was introduced in R2018b, so you'll have to linearize your array like I describe in my answer. doc for R2018a, doc for R2018b

Answer 2

[Rik]

This syntax should work for most releases:

sdev=std(m(:),'omitnan');