[Numpy-tickets] [NumPy] #341: Inconsistent behavior of comparison operators on character arrays, created with 'c' typecode

[NumPy]

#341: Inconsistent behavior of comparison operators on character arrays, created
with 'c' typecode
---------------------+------------------------------------------------------
 Reporter:  gkoczyk  |       Owner:  somebody
     Type:  defect   |      Status:  new     
 Priority:  normal   |   Milestone:          
Component:  Other    |     Version:  devel   
 Severity:  normal   |    Keywords:          
---------------------+------------------------------------------------------
 Arrays created using old Numeric 'c' typecode, seem to have '|S1'
 datatype, but raise errors on comparison with Python strings or sequences.


 {{{
 >>> numpy.array([ 'X' ], 'c')
 array(['X'],
       dtype='|S1')
 >>> numpy.array([ 'X' ], 'c') == 'X'
 Traceback (most recent call last):
   File "<stdin>", line 1, in ?
 TypeError: invalid string data-types in comparison
 >>> numpy.array([ 'X', 'X' ], 'c') < ['X']
 Traceback (most recent call last):
   File "<stdin>", line 1, in ?
 TypeError: invalid string data-types in comparison
 }}}


 On a sidenote, functor objects implementing comparisons on '|S1' arrays
 raise NotImplemented :

 :

 {{{
 >>> numpy.equal( numpy.array([ 'X', 'X' ], 'c') , numpy.array( ['X', 'X'],
 'c' ) )
 NotImplemented
 >>> numpy.equal( numpy.array([ 'X', 'X' ], 'c') , numpy.array( ['X', 'X'],
 '|S1' ) )
 NotImplemented
 >>> numpy.equal( numpy.array([ 'X', 'X' ], '|S1') , numpy.array( ['X',
 'X'], '|S1' ) )
 NotImplemented
 }}}

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/341>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.