[Numpy-tickets] [NumPy] #450: Make a.min() not copy data

[NumPy]

#450: Make a.min() not copy data
------------------------------------------+---------------------------------
 Reporter:  goddard                       |        Owner:  somebody
     Type:  defect                        |       Status:  new     
 Priority:  low                           |    Milestone:          
Component:  Other                         |      Version:  1.0.1   
 Severity:  normal                        |   Resolution:          
 Keywords:  reduce, memory, optimization  |  
------------------------------------------+---------------------------------
Changes (by oliphant):

  * priority:  high => low
  * milestone:  1.0.2 Release =>

Comment:

 This is not really a bug but could be considered a feature request.   The
 problem arises because you get a discontiguous array with b=a[1:,1:,1:]
 which when raveled must create a copy to obtain a 1-d array.

 It would be possible to perform a min over each axis instead of raveling,
 but this would be slower than raveling for certain sizes of arrays.  So,
 "fixing" this requires a bit more thinking then anybody has offered.

 My suggestion is to use

 a.min(axis=0).min(axis=0).min(axis=0)

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/450#comment:3>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.