Question

I would like to (quickly) determine if one array contains all the elements of another array, taking into account that the arrays may have repeated elements.

Thus, I tried something like this:

alice = %w(a a a b)
bob = %w(a a b c d e)
alice & bob => ["a", "b"]
alice - bob => []

But what I would like is an operator that will let me determine that bob does not include all the elements of alice, because bob doesn’t have enough “a” characters.

Answer 1

alice.select{|x| alice.count(x) > bob.count(x)}

Update
Setting up a benchmark:

require 'benchmark'

def subset_multivalue?(a, b)
  bb = b.clone
  a.each do |e|
    i = bb.index(e)
    if i
      bb.delete_at(i)
    else
      return false
    end
  end
  return true
end

def subset_multivalue2?(a, b)
  a.find{|x| a.count(x) > b.count(x)}
end

def subset_multivalue3?(alice, bob)
  alice_counts = alice.each_with_object(Hash.new(0)) { |o, h| h[o] += 1 }
  bob_counts = bob.each_with_object(Hash.new(0)) { |o, h| h[o] += 1 }

  alice_counts.all? do |k, v|
    bob_counts.has_key?(k) && bob_counts[k] >= v
  end
end

alice = %w(a a a b)
bob = %w(a a b c d e)

Benchmark.bm do |x|
  x.report("dave:") do
    1000000.times do
      subset_multivalue?(alice, bob)
    end
  end

  x.report("me:") do
    1000000.times do
      subset_multivalue2?(alice,bob)
    end
  end

  x.report("andrew:") do
    1000000.times do
      subset_multivalue3?(alice,bob)
    end
  end
end

Results:

       user     system      total        real
dave: 15.054000   0.000000  15.054000 ( 15.108864)
me: 11.529000   0.031000  11.560000 ( 11.689669)
andrew: 65.036000   0.047000  65.083000 ( 67.463859)