nan as output of partial correlation analysis

I'm running a partial correlation analysis on series of data.

How should I interpret a RHO=NaN?

Is there enything wrong or NaN is one of the possible results?

It occurs just for certain data sets.

Answer 1

[Oleg Komarov]

X = rand(100,2);
X(10) = NaN;
partialcorr(X)

You have NaNs in your data.

partialcorr(X,'rows','pairwise')

"A 'pairwise' value for the rows parameter can produce a RHO that is not positive definite. A 'complete' value always produces a positive definite RHO, but when data is missing, the estimates will be based on fewer observations, in general."