使用正则表达式从除撇号之外的字符串中去除所有字符和标点符号

Question

I attempted to let this method call:

alternate_words(". . . .  don’t let this stop you")

return every other word in the string, less punctuations except for ' .

This is the method definition:

def alternate_words(sentence)
  sentence.gsub(/[^a-z0-9\s']/i, "").split(" ").delete_if.with_index 
  {|word,index| index.odd? }
end

The result is:

["dont", "this", "you"]

The correct words are returned, but no ' is included. Changing the regex to:

/[^a-z0-9\s][']/i

returns

[".", ".", "don’t", "this", "you"]

Now, it correctly recognizes the apostrophe, but it incorrectly includes the periods. I don't understand why.

Answer 1

You may actually match words with apostrophes and hyphens with scan :

def alternate_words(sentence)
  sentence.scan(/[[:alnum:]]+(?:[’'-][[:alnum:]]+)*/).delete_if.with_index { |_,index| 
    index.odd? 
  }
end

p alternate_words(". . . . .  don’t let this stop you")
# => ["don’t", "this", "you"]

See a Ruby demo

The [[:alnum:]]+(?:[''-][[:alnum:]]+)* pattern may be enclosed with a word boundary - \\b - if you want to only match whole word.

Details :

• [[:alnum:]]+ - 1 or more alphanumeric symbols
• (?:[''-][[:alnum:]]+)* - zero or more (due to * , replace with another quantifier as per requirements) occurrences of:
  • [''-] - an apostrophe or a hyphen (the list may be adjusted_
  • [[:alnum:]]+ - 1 or more alphanumeric symbols.