[Numpy-tickets] [NumPy] #410: setmember1d problem

[NumPy]

#410: setmember1d problem
------------------------+---------------------------------------------------
 Reporter:  curiousjan  |        Owner:  somebody
     Type:  defect      |       Status:  closed  
 Priority:  normal      |    Milestone:          
Component:  Other       |      Version:          
 Severity:  normal      |   Resolution:  fixed   
 Keywords:              |  
------------------------+---------------------------------------------------
Changes (by rkern):

  * status:  new => closed
  * resolution:  => fixed

Old description:

> Please see the attached scripts.
> I am simply comparing two arrays, storing items that appear in both
> lists.
> I do this by
> a) storing all items from list 1 that are also in list 2
> b) storing all items from list 2 that are also in list 1
> The two resulting arrays should be identical, but they are not.
> A simple fix is
>
> svn diff numpy/lib
> Index: numpy/lib/arraysetops.py
> ===================================================================
> --- numpy/lib/arraysetops.py    (revision 3491)
> +++ numpy/lib/arraysetops.py    (working copy)
> @@ -126,7 +126,7 @@
>      zlike = nm.zeros_like
>      ar = nm.concatenate( (ar1, ar2 ) )
>      tt = nm.concatenate( (zlike( ar1 ), zlike( ar2 ) + 1) )
> -    perm = ar.argsort()
> +    perm = ar.argsort(kind='mergesort')
>      aux = ar[perm]
>      aux2 = tt[perm]
>  #    flag = ediff1d( aux, 1 ) == 0

New description:

 Please see the attached scripts.
 I am simply comparing two arrays, storing items that appear in both lists.
 I do this by
 a) storing all items from list 1 that are also in list 2
 b) storing all items from list 2 that are also in list 1
 The two resulting arrays should be identical, but they are not.
 A simple fix is

 svn diff numpy/lib

 {{{
 Index: numpy/lib/arraysetops.py
 ===================================================================
 --- numpy/lib/arraysetops.py    (revision 3491)
 +++ numpy/lib/arraysetops.py    (working copy)
 @@ -126,7 +126,7 @@
      zlike = nm.zeros_like
      ar = nm.concatenate( (ar1, ar2 ) )
      tt = nm.concatenate( (zlike( ar1 ), zlike( ar2 ) + 1) )
 -    perm = ar.argsort()
 +    perm = ar.argsort(kind='mergesort')
      aux = ar[perm]
      aux2 = tt[perm]
  #    flag = ediff1d( aux, 1 ) == 0
 }}}

Comment:

 Well, I'm applying your change anyways. It seems to me like a stable sort
 actually is required even if unique elements are provided. Good catch.

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/410#comment:4>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.