numarray question

[Colin J. Williams]

Jean Brouwers wrote:
> 
> You probably meant
> 
>   xyz = lambda x,y,z: float(x*y*z)
> 
> or maybe even
> 
> xyz = lambda x,y,z: float(x)*float(y)*float(z)
> 
> 
> /Jean Brouwers
> 
> 
Jean,

You are right, I overlooked SunX's wish to have a Float array.

Unfortunately, numarray assumes that the source is based in
the function indices, which delivers an Int array.

Thus, your suggestion raises an exception.  An alternative
is given below.

Colin W.

import numarray
xyz= lambda x,y,z: x*y*z
# This raises an exception
##xur = numarray.fromfunction(xyz, (6, 4, 3))
##  File "C:\Python23\Lib\site-packages\numarray\numarraycore.py", line 
715, in __float__
##    raise TypeError, "Only rank-0 numarray can be cast to floats."
# This gets around that problem
ind= numarray.indices(shape= (6, 4, 3), type= numarray.Float)
xur= apply(xyz, ind)
print xur




> 
> In article <2fCKc.1152$Vw3.217821 at news20.bellglobal.com>, Colin J.
> Williams <cjw at sympatico.ca> wrote:
> 
> 
>>SunX wrote:
>>
>>>I tried to initialize a float point array by:
>>>
>>>import numarray
>>>xur = numarray.fromfunction(lambda x,y,z:x*y*z, (2, 2, 2))
>>>
>>>but I ended up with an integer array even though x, y, and z are all floats.
>>>
>>>BTW, how do you unzip NumTut in windows?  And is there a newer version?
>>>
>>>Thank you
>>>Howard
>>
>>You might try something like:
>>
>>import numarray
>>xyz= lambda x,y,z:x*y*z
>>xur = numarray.fromfunction(xyz, (6, 6, 3))
>>print xur
>>