Should complex number have an ordering at all? (Trac #1406) #2004
Labels
Comments
|
@pv wrote on 2010-02-23 Numpy defines an lexicographical ordering (sort by real part first, then by imaginary part) for complex numbers, so the behavior of amax is in line with that. Changing that at this point requires a heavy reason to do so. |
|
Title changed from |
|
@charris wrote on 2010-05-25 Probably not, but since we have one there is no reason to toss it out, just don't use it if it doesn't do what is needed. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Original ticket http://projects.scipy.org/numpy/ticket/1406 on 2010-02-23 by trac user malev, assigned to unknown.
Hi there!
If you have and array with complex values, and try to execute: numpy.amax(array), amax, brings a result the max number of the array, but it only search in the real part.
I think that there shouldn't be result, like it happens when you call the max methods on a simple list.
In [1]: import numpy as np
In [2]: a_list = [1+2j, 1+3j, 2+2j,1+100j, 5+1j]
In [3]: a_np = np.array([1+2j, 1+3j, 2+2j,1+100j, 5+1j])
In [4]: np.amax(a_np)
Out[4]: (5+1j)
In [5]: max(a_list)
TypeError Traceback (most recent call last)
/home/malev/ in ()
TypeError: no ordering relation is defined for complex numbers
The code here: http://pastebin.com/b8JgMNhm
finally, I'm using:
Python 2.6.4 (r264:75706, Dec 7 2009, 18:45:15)
and numpy: '1.3.0'
Regards
malev
The text was updated successfully, but these errors were encountered: