更喜欢 %w(...) 而不是文字数组? [英] Prefer %w(...) to a literal array?
问题描述
在向 RubyMine 输入一个简单的字符串文字数组时:
While entering a simple literal array of strings in to RubyMine:
cols =
[
"Col1",
"Col2"
]
RubyMine 抱怨我更喜欢使用 %w
而不是文字数组:
RubyMine complained that I should prefer to use %w
over a literal array:
cols = %w(Col1 Col2)
它提供了一个指向 Ruby 样式指南的链接(此处),其中说:
It presented a link to a Ruby Style Guide (here), which said:
当你需要一个数组时,更喜欢 %w 而不是文字数组语法字符串.
Prefer %w to the literal array syntax when you need an array of strings.
# bad
STATES = ['draft', 'open', 'closed']
# good
STATES = %w(draft open closed)
我可以看到 %w
如何提供可能更简洁的代码.除了简洁之外,是否有任何理由更喜欢一种方法而不是另一种方法?
I can see how the %w
provides potentially more concise code. Aside from brevity, is there any reason to prefer one method over the other?
推荐答案
瞧!,一个基准:
require 'benchmark'
n = 1_000_000
Benchmark.bm(11) do |b|
b.report('%w') { n.times { %w[a b c d e f g h i j k l m n o p q r s t u v w x y z] } }
b.report('explicit') { n.times { ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z'] } }
b.report('numerics') { n.times { [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26] } }
end
user system total real
%w 2.590000 0.000000 2.590000 ( 2.591225)
explicit 2.590000 0.000000 2.590000 ( 2.584781)
numerics 0.300000 0.000000 0.300000 ( 0.309161)
user system total real
%w 2.590000 0.000000 2.590000 ( 2.591516)
explicit 2.590000 0.000000 2.590000 ( 2.584155)
numerics 0.300000 0.000000 0.300000 ( 0.308896)
user system total real
%w 2.590000 0.000000 2.590000 ( 2.592848)
explicit 2.590000 0.000000 2.590000 ( 2.585558)
numerics 0.300000 0.000000 0.300000 ( 0.308570)
我添加了数字"数组测试,因为我怀疑 %w
由于测试字符串性而比使用显式字符串更快.%w
不需要这样做,因为它假设一切都是字符串.运行三遍后,就处理字符串而言,这是一次清洗.数字规则、字符串流口水等等.
I added the "numerics" array test because I suspected that %w
is faster than using explicit strings due to testing for stringiness. %w
doesn't need to do that, because it assumes everything is a string. After running it three times it's a wash as far as dealing with strings. Numbers rule, strings drool, and all that.
之前的基准测试是在我的系统上使用 Ruby 1.9.3-p286 运行的.我在家里用我的旧 MacBook Pro 再次测试,使用 Ruby 1.8.7-p358,因此由于托管硬件的差异,加上运行较旧的 Ruby,以下数字较慢:
The previous benchmarks were run using Ruby 1.9.3-p286 on my system at work. I tested again using my old MacBook Pro at home, using Ruby 1.8.7-p358, so the following numbers are slower due to the differences in the hosting hardware, plus running an older Ruby:
user system total real
%w 3.070000 0.000000 3.070000 ( 3.080983)
explicit 3.100000 0.000000 3.100000 ( 3.093083)
numerics 0.950000 0.040000 0.990000 ( 0.990535)
user system total real
%w 3.080000 0.010000 3.090000 ( 3.076787)
explicit 3.090000 0.000000 3.090000 ( 3.089246)
numerics 0.950000 0.030000 0.980000 ( 0.989579)
user system total real
%w 3.080000 0.000000 3.080000 ( 3.073877)
explicit 3.090000 0.000000 3.090000 ( 3.091576)
numerics 0.950000 0.030000 0.980000 ( 0.989132)
在 1.8.7 上,%w
始终比之前快一点,这可能引起了速度谣言.
On 1.8.7, %w
was a tiny bit faster consistently, which probably gave rise to the speed-rumors.
这篇关于更喜欢 %w(...) 而不是文字数组?的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!