删除单词中除单个撇号和连字符之外的所有标点符号

Question

我之前问过一个类似的问题，但这个问题要具体得多，并且需要与之前提供的解决方案不同的解决方案，所以我希望可以发布它。我只需要在文本中保留撇号和字内破折号（删除所有其他标点符号）。比如我想从str1中获取str2：

str1<-"I'm dash before word -word, dash &%$,. in-between word, two before word --word just dashes ------, between words word - word"
str2<-"I'm dash before word word dash in-between word two before word  word just dashes  between words word  word"

到目前为止我的解决方案，首先删除单词之间的破折号：
gsub(" - ", " ", str1)

然后留下字母和数字字符以及剩余的破折号
gsub("[^[:alnum:]['-]", " ", str1)

问题是，它不会删除彼此跟随的破折号，例如“--”在单词的开头和结尾加上破折号：“-word”或“word--”

Answer 1

我认为这样做：

gsub('( |^)-+|-+( |$)', '\\1', gsub("[^ [:alnum:]'-]", '', str1))
#[1] "I'm dash before word word dash  in-between word two before word word just dashes  between words word  word"

Answer 2

这是一种方法：

gsub("([[:alnum:]][[:punct:]][[:alnum:]])|[[:punct:]]", "\\1", str1)
# [1] "I'm dash before word word dash  in-between word two before word word just dashes  between words word  word"

或者，更明确地说：

gsub("([[:alnum:]]['-][[:alnum:]])|[[:punct:]]", "\\1", str1)

同样的东西，略有不同/更短：

gsub("(\\w['-]\\w)|[[:punct:]]", "\\1", str1, perl=TRUE)

Answer 3

我建议

x <- "I'm dash before word -word, dash &%$,. in-between word, two before word --word just dashes ------, between words word - word"
gsub("\\b([-'])\\b|[[:punct:]]+", "\\1", x, perl=TRUE)
# =>  "I'm dash before word word dash  in-between word two before word word just dashes  between words word  word"

请参阅R demo。正则表达式是

\b([-'])\b|[[:punct:]]+

请参阅regex demo。详情：

• \b([-'])\b - - 或 ' 用单词字符（字母、数字或 _）括起来（注意：如果您只想保留在字母之间，请改用 (?<=\p{L})([-'])(?=\p{L})）
• | - 或
• [[:punct:]]+ - 1 个或多个标点符号。

要删除此替换后产生的任何前导/尾随和双空格字符，您可以使用

res <- gsub("\\b([-'])\\b|[[:punct:]]+", "\\1", x, perl=TRUE)
res <- trimws(gsub("\\s{2,}", " ", res))