Error using nanstd function

B=nanstd(X,0,'all');

when using the above statement in Matlab R2019a there are no problems, however when using it on another computer with R2017a, i get the error:

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 nanvar (line 32)

y = var(varargin{:},'omitnan');

Error in nanstd (line 28)

y = sqrt(nanvar(varargin{:}));

Answer 1

[Adam Danz]

If you're computing the std over all elements of X, this works in any release.

B = std(X(~isnan(X))); 

[Adam Evenoff]

That works, thanks!

Answer 2

[Steven Lord]

The ability to specify 'all' as a dimension in the var function (which the nanstd function calls) was introduced in release R2018b.

[Adam Evenoff]

Is there any other tool/way on R2017a to calculate standard deviation while omitting NaN values?

[Steven Lord]

The 'omitnan' option was introduced in release R2015a.

M = magic(5);
M(3, 3) = NaN;
std(M(:), 'omitnan')