It'll work, it is equivalent to the suggestion I made in my previous post with the f_inplace wrapper function (and it has the same drawback that numpy will allocate temporary memory, which wouldn't be the case if f was working in-place directly, by implementing it as "arr *= 2").

Note that you don't need to write x[:] * 2, you can write x * 2 directly.

-=- Olivier

2011/10/1 David Froger <>

Thanks everybody for the different solutions proposed, I really appreciate.
What about this solution? So simple that I didn't think to it...

import numpy as np
from numpy import *

def f(arr):
    return arr*2

a = array( [1,1,1] )
b = array( [2,2,2] )
c = array( [3,3,3] )
d = array( [4,4,4] )

for x in (a,b,c,d):
   x[:] = x[:]*2
   #instead of: x = x*2

print a
print b
print c
print d
NumPy-Discussion mailing list