如何在R语言中移除最外层大括号及前后所有内容？

输入字符串：

"A fancy ! nametag here! {{These} are the contents!}"

期望提取结果：

"{These} are the contents!"

要求：最外层大括号内部允许嵌套大括号，希望用正则表达式替代逐字符遍历的方案（原方案存在速度慢、占用内存、非向量化等问题）。

library(dplyr)
#This is the input
"A fancy ! nametag here! {{These} are the contents!}" %>%
  #Let's split them to a vector, so we can iterate character by character
  strsplit(split = "") %>%
  #Strsplit is vectorized, and returns a list as it should, but we need a vector (the first element of that list) at the moment
  unlist %>%
  {
    #We need to track whether we are inside parentheses
    insideParentheses = 0
    #Here we accrue the name tag
    nametag = ""
    #Here we accrue the contens
    contents = ""
    #And then we iterate character by character...
    for(char in .){
      #Checking and marking up whether we go into the next level of parentheses or come out
      insideParentheses = insideParentheses + (char=="{") - (char=="}")
      #If we are completely outside parentheses, we add the character to the name tag
      if(!insideParentheses) {nametag = paste0(nametag,char)}
      #If we have entered the most outer parentheses, but haven't got out yet, we write a character to contents
      if(insideParentheses) {contents = paste0(contents,char)}
    }
    
    #After each and every character have been iterated, we create a list
    list(
      contents %>% sub("\\{","",.) #Since the very first parenthesis is written to the contents, it needs to be removed
      ) %>%
      setNames(nametag %>% sub("\\}$", "", .)) #Since the very last parenthesis is written to the name tag, it needs to be removed too
  } %>%
  print

运行结果：

$`A fancy ! nametag here! `
[1] "{These} are the contents!"

R的PCRE正则支持递归匹配，可完美处理嵌套大括号场景，且是向量化操作，性能远优于逐字符遍历。

核心提取代码 #

# 基础R版本
input <- "A fancy ! nametag here! {{These} are the contents!}"
# 匹配最外层大括号及内部所有内容（含嵌套）
matched <- regmatches(input, regexpr("\\{(?:[^{}]|(?R))+\\}", input, perl = TRUE))
# 去掉最外层一对大括号，得到目标结果
final_result <- sub("^\\{(.*)\\}$", "\\1", matched)
print(final_result)

输出：

[1] "{These} are the contents!"

正则表达式解释 #

• \\{：匹配最外层左大括号
• (?:[^{}]|(?R))+：非捕获组，重复匹配两种情况：
  • [^{}]：匹配非大括号的任意字符
  • (?R)：递归调用整个正则，处理嵌套的大括号结构
• \\}：匹配最外层右大括号

复刻原代码的列表输出 #

如果需要和原代码一样返回带名称的列表：

library(stringr)
input <- "A fancy ! nametag here! {{These} are the contents!}"

# 提取标签部分（大括号外的内容）
nametag <- str_remove(input, "\\{(?:[^{}]|(?R))+\\}$")
# 提取内容并处理格式
contents <- str_extract(input, "\\{(?:[^{}]|(?R))+\\}") %>% 
  str_remove("^\\{") %>% 
  str_remove("\\}$")

# 生成结果列表
result_list <- setNames(list(contents), nametag)
print(result_list)

运行结果与原代码完全一致：

$`A fancy ! nametag here! `
[1] "{These} are the contents!"

正则方案是原生向量化操作，无需循环遍历每个字符，处理大量字符串时速度更快、内存占用更低，完全符合R的向量化编程范式。

内容的提问来源于stack exchange，提问作者Aku-Ville Lehtimäki