[Numpy-tickets] [NumPy] #241: maximum and nan / identity for maximum

NumPy numpy-tickets at scipy.net
Fri Aug 11 09:32:56 CDT 2006


#241: maximum and nan / identity for maximum
-----------------------------+----------------------------------------------
 Reporter:  martin_wiechert  |       Owner:  somebody   
     Type:  enhancement      |      Status:  new        
 Priority:  normal           |   Milestone:  1.0 Release
Component:  numpy.core       |     Version:  devel      
 Severity:  normal           |    Keywords:             
-----------------------------+----------------------------------------------
 How should maximum handle nans? I think there are only two reasonable
 behaviours. Either nans should be ignored or the result should be nan.
 (I personally favour a keyword switch over a separate function nanmax.)

 In the current implementation the outcome depends on the ''order'' of
 elements:
 {{{
 >>> import numpy
 >>> numpy.version.version
 '1.0b1'
 >>> from numpy import *
 >>> a = array ((1, nan))
 >>> maximum.reduce (a)
 1.0
 >>> maximum.reduce (a [::-1])
 nan
 }}}
 This is certainly confusing if not a bug.

 As a related request I'd like to see an identity element for maximum,
 namely -inf.

 All the above of course also holds for minimum.

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/241>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list