Setting a complex variable (or array) to NaN or np.nan gives nan + 0j, and the imaginary part of this is not NaN.
@mabitbol Let's change the fill value for bad or missing samples to nan + 1j * nan (or nan * (1 + 1j), whatever) so that calculations don't accidentally treat the imaginary part of these samples as good data.
Setting a complex variable (or array) to NaN or np.nan gives nan + 0j, and the imaginary part of this is not NaN.
@mabitbol Let's change the fill value for bad or missing samples to nan + 1j * nan (or nan * (1 + 1j), whatever) so that calculations don't accidentally treat the imaginary part of these samples as good data.