Fast Ruby for hash_update_vs_hash_brackets

[manjunath724]

Hash#update is an alias for Hash#merge!

Currently we have merge_bang_vs_[]= which suggests that Hash#[]= is faster and Hash#merge!() is slower. But we dont have any suggestions for update_vs_[]= i.e., Hash#update() vs Hash#[]=.

require 'benchmark/ips'

ENUM = (1..100)

def fast
  ENUM.each_with_object({}) do |e, h|
    h[e] = e
  end
end

def slow
  ENUM.each_with_object({}) do |e, h|
    h.update(e => e)
  end
end

Benchmark.ips do |x|
  x.report('Hash#[]=') { fast }
  x.report('Hash#update') { slow }
  x.compare!
end

##### `Hash#update` vs `Hash#[]=` [code](code/hash/update-vs-\[\]=.rb)

$ ruby -v code/hash/update-vs-\[\]=.rb
ruby 2.6.6p146 (2020-03-31 revision 67876) [x86_64-darwin18]

Warming up --------------------------------------
            Hash#[]=     7.453k i/100ms
         Hash#update     4.311k i/100ms
Calculating -------------------------------------
            Hash#[]=     74.764k (± 1.9%) i/s -    380.103k in   5.085962s
         Hash#update     43.220k (± 0.8%) i/s -    219.861k in   5.087364s

Comparison:
            Hash#[]=:    74764.0 i/s
         Hash#update:    43220.1 i/s - 1.73x  (± 0.00) slower

Can we extend this to Hash#update()?