我该如何排序这个数组？

Question

我有一个数组，headlines，保存几个句子，所以像：

headlines = ["I see a tree", "Facebook is slow", "plants need water to grow", "There's an orange", "I think we'll agree"] 
first = headlines[0] 
second = headlines[1] 
third = headlines[2] 

我使用ruby_rhymes宝石所提供的方法#to_phrase.rhymes打印出押韵的最后一个字的话在你提供一个字符串用。我们检查字符串数组韵，我做这样的事情：

> first.to_phrase.rhymes.flatten.join(", ").include?(second.to_phrase.rhymes.flatten.join(", ")) 
=> false 
> second.to_phrase.rhymes.flatten.join(", ").include?(third.to_phrase.rhymes.flatten.join(", ")) 
=> true 

我希望将这些保存到一个文本文件，所以我想他们在阵列中进行排序，以便押韵对是彼此以后。我知道要排序，所以如果最后3个字符是相同的字符串是：

headlines.sort! {|a,b| a[-3,3] <=> b[-3,3] } 

但我不知道如何去做我想要的。

Answer 1

所以我想通了：

headlines.sort_by! { |h| h.to_phrase.rhyme_key } 

这不起作用100％，但是这的故障宝石依赖的字典。

Answer 2

通过研究你的建议的输出，你可以看到你在正确的轨道上：

p headlines.sort {|a,b| a[-3,3] <=> b[-3,3] } 
# => ["Facebook is slow", "There's an orange", "I see a tree", "I think we'll agree", "plants need water to grow"] 

“......慢”和“...成长”是唯一的无序的句子，引起字母'r'和'o'。一个简单的黑客将扭转这样的比较顺序：

p headlines.sort {|a,b| a[-3,3].reverse <=> b[-3,3].reverse } 
# => ["I see a tree", "I think we'll agree", "There's an orange", "Facebook is slow", "plants need water to grow"]