I noticed that numpy.linalg.eigvalsh returns a complex array, even though mathematically the resulting eigenvalues are guaranteed to be real.
Looking at the source code, the underlying zheevd routine of LAPACK indeed returns an array of real numbers which is than converted to complex in the numpy wrapper. Does numpy policy require the type of the result to be the same as the type of input? Copying an array twice to arrive at the original result seems pointless to me. _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion
