U_REGEX_INVALID_CAPTURE_GROUP_NAME error occurs wh

2019-02-25 14:26发布

问题:

I recently implemented a function to escape characters interpretable as regex that go into a system call for my R package 'rNOMADS'

SanitizeWGrib2Inputs <- function(check.strs) {
    #Escape regex characters before inputting to wgrib2
    #INPUTS
    #    CHECK.STRS - Strings possibly containing regex metacharacters
    #OUTPUTS
    #    CHECKED.STRS - Strings with metacharacters appropriately escaped

    meta.chars <- paste0("\\", c("(", ")", ".", "+", "*", "^", "$", "?", "[", "]", "|"))

   for(k in 1:length(meta.chars)) {
       check.strs <- stringr::str_replace(check.strs, meta.chars[k], paste0("\\\\", meta.chars[k]))
   }

   checked.strs <- check.strs

   return(checked.strs)
}

and I include an example in my package documentation:

check.strs <- c("frank", "je.rry", "har\\old", "Johnny Ca$h")
checked.strs <- SanitizeWGrib2Inputs(check.strs) 

This works fine on my ubuntu machine and passed the CRAN checks. However, when I uploaded the package to CRAN, their windows checker said:

> check.strs <- c("frank", "je.rry", "har\\old", "Johnny Ca$h")
> checked.strs <- SanitizeWGrib2Inputs(check.strs) Error in stri_replace_first_regex(string, pattern, fix_replacement(replacement),  :    Invalid capture group name. (U_REGEX_INVALID_CAPTURE_GROUP_NAME) Calls: SanitizeWGrib2Inputs -> <Anonymous> -> stri_replace_first_regex -> .Call Execution halted
** running examples for arch 'x64' ... ERROR Running examples in 'rNOMADS-Ex.R' failed The error most likely occurred in:

> base::assign(".ptime", proc.time(), pos = "CheckExEnv")
> ### Name: SanitizeWGrib2Inputs
> ### Title: Make sure regex metacharacters are properly escaped
> ### Aliases: SanitizeWGrib2Inputs
> ### Keywords: internal
> 
> ### ** Examples
> 
> 
> check.strs <- c("frank", "je.rry", "har\\old", "Johnny Ca$h")
> checked.strs <- SanitizeWGrib2Inputs(check.strs) Error in stri_replace_first_regex(string, pattern, fix_replacement(replacement),  :    Invalid capture group name. (U_REGEX_INVALID_CAPTURE_GROUP_NAME) Calls: SanitizeWGrib2Inputs -> <Anonymous> -> stri_replace_first_regex -> .Call Execution halted

I verified this behavior on my Windows partition. What is the work around for this?

回答1:

Use a simple str_replace_all to escape all the special regex metacharacters:

SanitizeWGrib2Inputs <- function(check.strs) {
    return(str_replace_all(check.strs, "[{\\[()|?$^*+.\\\\]", "\\$0"))
}
check.strs <- c("frank", "je.rry", "har\\old", "Johnny Ca$h")
checked.strs <- SanitizeWGrib2Inputs(check.strs) 
checked.strs
## => [1] "frank"         "je\\.rry"      "har\\\\old"    "Johnny Ca\\$h"

Notes:

  • The "[{\\[\\]()|?$^*+.\\\\] (actually, "[{\[\]()|?$^*+.\\]) will match any single char, either {, [, ], (, ), |, ?, $, ^, *, +, . or \
  • The "\\$0" replacement will change each of the chars into \ + the same char (the $0 is the backreference to the whole match value).
  • I do not think you need to add ] here since outside a character class it is not a special character if there is no paired opening [ in front.


标签: r regex escaping