Sample an array with seed value to get consistent result

Question

I am sampling several arrays and would like to add a seed value to get a consistent result every time its run, now and in the future.

My example:

constant_seed_value = 123456789
["a","b","c"].sample(seed: constant_seed_value ) should return "a" when run every time.

Answer 1

Just pass a Random.new with your seed to sample:

%w[a b c].sample(1, random: Random.new(123456789))
#=> ["a"]

See Array#sample and Random

Answer 2

You will need to make the seed a constant (or write it to a file and read it each time the program is run). I assume you don't care what the seed actually is.

If a side-calculation produces

Random.new_seed
  #=> 44220669194288225494276674522501875094 

you will write

SEED = 44220669194288225494276674522501875094

See Random::new_seed.

When the program is run you must initialize the seed to this value.

Random.srand(SEED)
  #=> 129123040985656142450143558000073927364

See Random::srand.

Now let's compute some pseudo-random values.

arr = (1..1000).to_a

arr.sample(4)
  #=> [762, 619, 41, 997] 
rand
  #=> 0.9619996498741139 
rand
  #=> 0.7952214967836931 

Now restore the initial seed, as we would do if we re-ran the program.

Random.srand(SEED)
  #=> 190782386885144604306184636344084340916 

If we repeat the initial constuction of pseudo-random values we see they are the same as when we first computed them.

arr.to_a.sample(4)
  #=> [762, 619, 41, 997]
rand
  #=> 0.9619996498741139 
rand
  #=> 0.7952214967836931