Regex R strsplit：根据字符分割，除非后面跟着特定字符

假设我有一个字符串向量，比如

split_these = c("File Location:C:\\Documents","File Location:Pete's Computer","File Location:") 

我想根据“：”分割此向量中的每个元素，但后面有“\”时除外。我想要的是回报的东西，比如

#preferred solution
"File Location" "C:\\Documents"
"File Location" "Pete's Computer"
"File Location" ""

或

我试过以下方法

strsplit(split_these, ":")
[[1]]
[1] "File Location" "C"             "\\Documents"  

[[2]]
[1] "File Location" "Pete Computer"

[[3]]
[1] "File Location"

strsplit(split_these, ":[^\\]")
[[1]]
[1] "File Location" ":\\Documents" 

[[2]]
[1] "File Location" "ete Computer" 

[[3]]
[1] "File Location:"

---

我建议使用带有否定前瞻断言的PCREs。还要注意，需要对反斜杠进行双转义，因为它在R字符串和正则表达式语法中都充当元字符

strsplit(perl=T,split_these,':(?!\\\\)');
## [[1]]
## [1] "File Location" "C:\\Documents"
##
## [[2]]
## [1] "File Location"   "Pete's Computer"
##
## [[3]]
## [1] "File Location"

如果要将列表简化为单个字符向量：

do.call(c,strsplit(perl=T,split_these,':(?!\\\\)'));
## [1] "File Location" "C:\\Documents" "File Location" "Pete's Computer" "File Location"

---

我想出了一个办法来获取尾随的空字符串字段。由于

strsplit（）

总是省略最后一个空字段，因此我们可以简单地将分隔符连接到每个输入字符串的末尾。如果原始字符串中没有尾随分隔符，则将忽略新的空字段，而不会更改结果。如果原始字符串中有尾随分隔符，则我们将获得所需的空字段：

do.call(c,strsplit(perl=T,paste0(split_these,':'),':(?!\\\\)'));
## [1] "File Location" "C:\\Documents" "File Location" "Pete's Computer" "File Location" ""

---

使用

read.dcf

对

split\u这些

元素进行迭代，可以得到一个命名的字符向量，该向量可以被重组为一个data.frame：

v <- drop(do.call("cbind", lapply(split_these, function(x) read.dcf(textConnection(x)))))

或

给予：

> v
    File Location     File Location     File Location 
  "C:\\Documents" "Pete's Computer"                "" 

            ind          values
1 File Location   C:\\Documents
2 File Location Pete's Computer
3 File Location

> stack(v)[2:1]

            ind          values
1 File Location   C:\\Documents
2 File Location Pete's Computer
3 File Location