[Numpy-discussion] subclassing float64 (and friends)

Travis Oliphant oliphant at ee.byu.edu
Fri Jan 5 08:23:35 CST 2007


Stefan van der Walt wrote:
> On Wed, Jan 03, 2007 at 05:12:40PM -0600, eric jones wrote:
>   
>> Thanks for the update.  For now, I'll try doing what I need to by 
>> sub-classing float.  But, I'm gonna miss __array_finalize__ :-).
>>     
>
> Looks like r3493 is intended to fix this.  The 'view' method works
> now, but I still see
>
> In [2]: import numpy as N
>
> In [3]: x = N.float64(12)
>
> In [4]: class Floaty(N.float64):
>    ...:     def __new__(subclass,data):
>    ...:         return data.view(subclass)
>
> In [6]: Floaty(x)
> Out[6]: 12.0
>
> In [7]: type(Floaty(x))
> Out[7]: <type 'numpy.float64'>
>
>   
I did not change the view method.  This will still always return 
base-class array scalars.   Could look into it.

There is no __array_finalize__, however, for array scalars.

-Travis



More information about the Numpy-discussion mailing list