Preferred Random Library

Robert Kern robert.kern at gmail.com
Thu Apr 26 07:56:50 CEST 2007


Bill Jackson wrote:
> Is there a preferred random library?
> 
> scipy.random
> random
> 
> Besides scipy's library returning ndarrays, is there any other 
> advantage/disadvantage?

numpy.random (since that's where scipy.random comes from, I recommend always
referring to it as numpy.random) can be a bit faster since it is implemented as
entirely an extension module; parts of random are implemented in Python and
incur Python function call overhead. It's certainly faster if you need a lot of
numbers at once. numpy.random has several more non-uniform distributions
implemented.

numpy.random does not implement jumpahead(). random, naturally, does not require
a third-party package.

By and large, I'd say the distinguishing factor is whether or not you want to
use numpy in your program. If so, then use numpy.random. If not, then use random.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco




More information about the Python-list mailing list