Question about 'random'

Dev Shed Forums Sponsor:

January 24th, 2004, 07:23 PM

theperfectsoup

Contributing User

Join Date: Jul 2003

Posts: 35

Time spent in forums: < 1 sec
Reputation Power: 10

Question about 'random'

If I have multiple files in a single project using function randint, should I run seed() somewhere in each file? Or can I just run seed() once and have it seed all randint calls across all files?

This question may seem silly, but I remember in C++ that the more you call seed, the less random your numbers are! If this is the case in Python, I'd like to avoid that effect...

Thanks,
tps

January 24th, 2004, 07:50 PM

netytan

Hello World :)

Join Date: Mar 2003

Location: Hull, UK

Posts: 2,537

Time spent in forums: 1 Week 2 Days 18 h 17 m 47 sec
Reputation Power: 68

I've used 'random' in alot of different applications and i have to say i've never once needed seed()

, all you really need is randint(), then let Python handle the rest

Code:

>>> import random
>>> random.randint(1, 10)
5
>>> random.randint(1, 10)
6
>>> random.randint(1, 10)
2
>>> random.randint(1, 10)
1
>>> random.randint(1, 10)
7
>>> random.randint(1, 10)
2
>>> random.randint(1, 10)
7
>>> random.randint(1, 10)
2
>>> random.randint(1, 10)
8
>>>

Let me know if you have any questions

Mark.

__________________
programming language development: www.netytan.com – Hula

January 24th, 2004, 11:50 PM

Strike

Contributing User

Join Date: Dec 2001

Location: Houston, TX

Posts: 383

Time spent in forums: 1 h 41 m 27 sec
Reputation Power: 12

seed() is for making reproducible results, really. Python uses the "Mersenne Twister" algorithm to produce random numbers, and while it is periodic, it has a HUGE period. In fact, I know the exact size: (2**19937)-1. So if you start at the same point, you actually will get the same set of values for each time after. Observe:

Code:

>>> import random
>>> random.randint(1,100)
36
>>> random.randint(1,100)
93
>>> random.seed(42)
>>> random.randint(1,100)
64
>>> random.randint(1,100)
3
>>> random.randint(1,100)
28
>>> random.seed(42)
>>> random.randint(1,100)
64
>>> random.randint(1,100)
3
>>> random.randint(1,100)
28

It's a way you can test and make sure you are getting the "expected" random value in automated tests, for example. But you never have to explicitly seed the generator, as it will automatically seed it with the system time, I believe.

__________________
Debian - because life's too short for worrying.
Best. (Python.) IRC bot. ever.

January 25th, 2004, 12:27 AM

Scorpions4ever

Banned ;)

Dev Shed God 9th Plane (9000 - 9499 posts)

Join Date: Nov 2001

Location: Woodland Hills, Los Angeles County, California, USA

Posts: 9,404

Time spent in forums: 2 Months 9 h 11 m 7 sec
Reputation Power: 4080

The random generation module automatically seeds itself with the current system time, when you first import the module. Also, another interesting thing to note is that if you call random.seed() in your code with no arguments (or None), it will use the current system time as the seed.
[edit]Note that older versions of python (below 2.1) used Wichmann Hill instead of Mersenne Twister as the algorithm for the random module.[/edit]

__________________
Up the Irons
What Would Jimi Do? Smash amps. Burn guitar. Take the groupies home.
"Death Before Dishonour, my Friends!!" - Bruce D ickinson, Iron Maiden Aug 20, 2005 @ OzzFest
Down with Sharon Osbourne

Last edited by Scorpions4ever : January 25th, 2004 at 12:37 AM.

Question about 'random'

Developer Shed Advertisers and Affiliates