benchmark demonstrating that #keys.each performs marginally better than #each_key

Author: JonRowe

benchmarks/keys_each_vs_each_key.rb

@@ -0,0 +1,43 @@
+require 'benchmark/ips'
+
+small_hash = { :key => true, :more_key => true, :other_key => true }
+large_hash = (1...100).inject({}) { |hash, key| hash["key_#{key}"] = true; hash }
+
+Benchmark.ips do |x|
+  x.report('keys.each with small hash') do
+    small_hash.keys.each { |value| value == true }
+  end
+
+  x.report('each_key with small hash') do
+    small_hash.each_key { |value| value == true }
+  end
+
+  x.report('keys.each with large hash') do
+    large_hash.keys.each { |value| value == true }
+  end
+
+  x.report('each_key with large hash') do
+    large_hash.each_key { |value| value == true }
+  end
+end
+
+__END__
+
+Calculating -------------------------------------
+keys.each with small hash
+                       105.581k i/100ms
+each_key with small hash
+                       112.045k i/100ms
+keys.each with large hash
+                         7.625k i/100ms
+each_key with large hash
+                         6.959k i/100ms
+-------------------------------------------------
+keys.each with small hash
+                          2.953M (± 3.8%) i/s -     14.781M
+each_key with small hash
+                          2.917M (± 4.0%) i/s -     14.678M
+keys.each with large hash
+                         79.349k (± 2.5%) i/s -    396.500k
+each_key with large hash
+                         72.080k (± 2.1%) i/s -    361.868k