Re: [Numpy-discussion] What should be the result in some statistics corner cases?