[Numpy-tickets] [NumPy] #549: int() of numpy scalar fails silently

NumPy numpy-tickets@scipy....
Wed Jul 11 15:09:22 CDT 2007


#549: int() of numpy scalar fails silently
------------------------+---------------------------------------------------
 Reporter:  zouave      |        Owner:  somebody     
     Type:  defect      |       Status:  new          
 Priority:  normal      |    Milestone:  1.0.4 Release
Component:  numpy.core  |      Version:  devel        
 Severity:  major       |   Resolution:               
 Keywords:              |  
------------------------+---------------------------------------------------
Comment (by zouave):

 I just added a patch that should solve most conversions of numarray
 scalars to python ints.

 e.g.
 {{{
 from numpy import *

 l= [1e6, 1e12, 1e18, -1e6, -1e12, -1e18]
 af64=  array(l, dtype='float64')
 ai64=  array(l, dtype='int64')
 aui64= array(l[:3], dtype='uint64')

 il=     map(int, l)
 iaf64=  map(int, af64)
 iai64=  map(int, ai64)
 iaui64= map(int, aui64)

 print il
 print iaf64
 print iai64
 print iaui64
 }}}

 Output before patch:
 {{{
 [1000000, 1000000000000L, 1000000000000000000L, -1000000, -1000000000000L,
 -1000000000000000000L]
 [1000000, -2147483648, -2147483648, -1000000, -2147483648, -2147483648]
 [1000000, -727379968, -1486618624, -1000000, 727379968, 1486618624]
 [1000000, -727379968, -1486618624]
 }}}

 Output after patch:
 {{{
 [1000000, 1000000000000L, 1000000000000000000L, -1000000, -1000000000000L,
 -1000000000000000000L]
 [1000000, 1000000000000L, 1000000000000000000L, -1000000, -1000000000000L,
 -1000000000000000000L]
 [1000000, 1000000000000L, 1000000000000000000L, -1000000, -1000000000000L,
 -1000000000000000000L]
 [1000000, 1000000000000L, 1000000000000000000L]
 }}}

 Does it look ok to numpy's gurus?  If so, can it be applied to the svn
 repository?  Or, can I commit this change myself?

 Thanks,

 Xavier

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/549#comment:2>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list