.Try vs && performance

Var*_*rac 4 ruby-on-rails

假设我有一个来自SQL查询的传入值,如下所示:

grok = Foo.select(:foo_attr1, :foo_attr2).first
Run Code Online (Sandbox Code Playgroud)

foo_attr2是一个可以为空的领域.现在假设我需要对输出进行处理(如果存在的话).

krug = grok.foo_attr2.try(:bar).try(:baz)
gnar = grok.foo_attr2 && grok.foo_attr2.bar.baz # Assumes bar will always return output that can be baz'd
Run Code Online (Sandbox Code Playgroud)

这两个操作中哪一个更好用,为什么?

Sur*_*rya 7

使用gnar = grok.foo_attr2 && grok.foo_attr2.bar.bazRuby的逻辑运算符完成后肯定会更快.虽然try是由Rails引入并进行额外的if-else条件检查.来自代码:

# File activesupport/lib/active_support/core_ext/object/try.rb, line 41
def try(*a, &b)
  if a.empty? && block_given?
    yield self
  else
    public_send(*a, &b) if respond_to?(a.first)
  end
end
Run Code Online (Sandbox Code Playgroud)

好吧,这是一个基准,以准确显示我想说的内容:

class Object
  def try(*a, &b)
    if a.empty? && block_given?
      yield self
    else
      public_send(*a, &b) if respond_to?(a.first)
    end
  end
end

class Foo
  attr_reader :a
  def initialize(a = nil)
    @a = a    
  end
end

require "benchmark"


bar = Foo.new
baz = Foo.new(1)

n = 10000000
Benchmark.bm(40) do |x|
  x.report("try"){ n.times { bar.a.try(:class).try(:to_s) } }
  x.report("&& "){ n.times { baz.a && baz.a.class.to_s } }
end
Run Code Online (Sandbox Code Playgroud)

结果是:

           user      system      total        real
try      10.800000  0.030000   10.830000  ( 10.829770)
&&        3.940000   0.010000   3.950000  (  3.944375)
Run Code Online (Sandbox Code Playgroud)