[NumPy-Tickets] [NumPy] #1608: numpy.sort on a flattened array has unwanted side-effect

[NumPy Trac]

#1608: numpy.sort on a flattened array has unwanted side-effect
---------------------------+------------------------------------------------
 Reporter:  carlscheffler  |       Owner:  somebody
     Type:  defect         |      Status:  new     
 Priority:  normal         |   Milestone:  1.5.1   
Component:  numpy.core     |     Version:  1.3.0   
 Keywords:  sort, flat     |  
---------------------------+------------------------------------------------
 (The code example below should be self-explanatory.)

 The sort function has the side-effect of changing/removing the view
 information of its argument when called with argument.flat. This side-
 effect does not appear when using sort(ravel(argument)) rather than
 sort(argument.flat).

 Tested on
  * Python version 2.6.5, NumPy version 1.3.0; and
  * Python version 2.6.2, NumPy version 1.2.1.

 {{{
 import numpy
 x = numpy.array([[1,2],[3,4]])
 x = x.transpose() # Create a transposed view of the original array
 print x           # [[1, 3], [2, 4]] -- OK
 y = numpy.sort(numpy.ravel(x))
 print x           # [[1, 3], [2, 4]] -- OK
 y = numpy.sort(x.flat)
 print x           # [[1, 2], [3, 4]] -- ERROR
 }}}

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1608>
NumPy <http://projects.scipy.org/numpy>
My example project