How to get a random number in Ruby

If you're not only seeking for a number but also hex or uuid it's worth mentioning that the SecureRandom module found its way from ActiveSupport to the ruby core in 1.9.2+. So without the need for a full blown framework:

require 'securerandom'

p SecureRandom.random_number(100) #=> 15
p SecureRandom.random_number(100) #=> 88

p SecureRandom.random_number #=> 0.596506046187744
p SecureRandom.random_number #=> 0.350621695741409

p SecureRandom.hex #=> "eb693ec8252cd630102fd0d0fb7c3485"

It's documented here: Ruby 1.9.3 - Module: SecureRandom (lib/securerandom.rb)

You can generate a random number with the rand method. The argument passed to the rand method should be an integer or a range, and returns a corresponding random number within the range:

rand(9)       # this generates a number between 0 to 8
rand(0 .. 9)  # this generates a number between 0 to 9
rand(1 .. 50) # this generates a number between 1 to 50
#rand(m .. n) # m is the start of the number range, n is the end of number range

While you can use rand(42-10) + 10 to get a random number between 10 and 42 (where 10 is inclusive and 42 exclusive), there's a better way since Ruby 1.9.3, where you are able to call:

rand(10...42) # => 13

Available for all versions of Ruby by requiring my backports gem.

Ruby 1.9.2 also introduced the Random class so you can create your own random number generator objects and has a nice API:

r = Random.new
r.rand(10...42) # => 22
r.bytes(3) # => "rnd"

The Random class itself acts as a random generator, so you call directly:

Random.rand(10...42) # => same as rand(10...42)

Notes on Random.new

In most cases, the simplest is to use rand or Random.rand. Creating a new random generator each time you want a random number is a really bad idea. If you do this, you will get the random properties of the initial seeding algorithm which are atrocious compared to the properties of the random generator itself.

If you use Random.new, you should thus call it as rarely as possible, for example once as MyApp::Random = Random.new and use it everywhere else.

The cases where Random.new is helpful are the following:

• you are writing a gem and don't want to interfere with the sequence of rand/Random.rand that the main programs might be relying on
• you want separate reproducible sequences of random numbers (say one per thread)
• you want to be able to save and resume a reproducible sequence of random numbers (easy as Random objects can marshalled)

Use rand(range)

From Ruby Random Numbers:

If you needed a random integer to simulate a roll of a six-sided die, you'd use: 1 + rand(6). A roll in craps could be simulated with 2 + rand(6) + rand(6).

Finally, if you just need a random float, just call rand with no arguments.

---

As Marc-André Lafortune mentions in his answer below (go upvote it), Ruby 1.9.2 has its own Random class (that Marc-André himself helped to debug, hence the 1.9.2 target for that feature).

For instance, in this game where you need to guess 10 numbers, you can initialize them with:

10.times.map{ 20 + Random.rand(11) } 
#=> [26, 26, 22, 20, 30, 26, 23, 23, 25, 22]

Note:

• Using Random.new.rand(20..30) (using Random.new) generally would not be a good idea, as explained in detail (again) by Marc-André Lafortune, in his answer (again).

• But if you don't use Random.new, then the class method rand only takes a max value, not a Range, as banister (energetically) points out in the comment (and as documented in the docs for Random). Only the instance method can take a Range, as illustrated by generate a random number with 7 digits.

This is why the equivalent of Random.new.rand(20..30) would be 20 + Random.rand(11), since Random.rand(int) returns “a random integer greater than or equal to zero and less than the argument.” 20..30 includes 30, I need to come up with a random number between 0 and 11, excluding 11.