正则表达式拆分字符串，不包括可转义引号之间的分隔符

Question

除非空格包含在可转义的引号中，否则我需要能够用空格分隔字符串。换句话说，spam spam spam "and \"eggs" 应该返回 spam、spam、spam 和 and "eggs。我打算在 python 中使用re.split 方法来执行此操作，您可以在其中使用正则表达式识别要拆分的字符。

我发现这个可以找到可转义引号之间的所有内容：

((?<![\\])['"])((?:.(?!(?<![\\])\1))*.?)\1

来自：https://www.metaltoad.com/blog/regex-quoted-string-escapable-quotes 并且 this 按字符拆分，除非在引号之间：

\s(?=(?:[^\"]*\"[^\"]*\")*[^\"]*$)

来自：https://stackabuse.com/regex-splitting-by-character-unless-in-quotes/。这会查找在空格和行尾之间有偶数个双引号的所有空格。

我正在努力将这两个解决方案结合在一起。

对于参考参考，我发现了这个我发现了这个超级有用的正则表达式备忘单：https://medium.com/factory-mind/regex-tutorial-a-simple-cheatsheet-by-examples-649dc1c3f285

我还发现https://regex101.com/ 非常有用：允许您测试正则表达式

Answer 1

终于成功了：

\s(?=(?:(?:\\\"|[^\"])*(?<!\\)\"(?:\\\"|[^\"])*(?<!\\)\")*(?:\\\"|[^\"])*$)

这结合了问题中的两个解决方案，以在右侧找到偶数个未转义双引号的空格。说明：

\s                 # space
(?=                # followed by (not included in match though)
  (?:              # match pattern (but don't capture)
    (?:
      \\\"         # match escaped double quotes
      |            # OR
      [^\"]        # any character that is not double quotes
    )*             # 0 or more times
    (?<!\\)\"      # followed by unescaped quotes
    (?:\\\"|[^\"])*  # as above match escaped double quotes OR any character that is not double quotes
    (?<!\\)\"      # as above - followed by unescaped quotes
                   # the above pairs of unescaped quotes
  )*               # repeated 0 or more times (acting on pairs of quotes given an even number of quotes returned)
  (?:\\\"|[^\"])*  # as above
  $                # end of the line
)

所以最终的python是：

import re
test_str = r'spam spam spam "and \"eggs"'
regex = r'\s(?=(?:(?:\\\"|[^\"])*(?<!\\)\"(?:\\\"|[^\"])*(?<!\\)\")*(?:\\\"|[^\"])*$)'

test_list = re.split(regex, test_str)
print(test_list)
>>> ['spam', 'spam', 'spam', '"and \\"eggs"']

这种方法唯一的缺点是它会留下前导的尾随引号，但是我可以使用以下 python 轻松识别和删除它们：

# remove leading and trailing unescaped quotes
test_list = list(map(lambda x: re.sub(r'(?<!\\)"', '', x), test_list))
# remove escape characters - they are no longer required
test_list = list(map(lambda x: x.replace(r'\"', '"'), test_list))
print(test_list)
>>> ['spam', 'spam', 'spam', 'and "eggs']