CODEWITHSUNDEEP
windowsrfilenames
DMcMor
DMcMor

Reputation: 183

How can I check if a string is a valid file name for windows using R?

I've been writing a program in R that outputs randomization schemes for a research project I'm working on with a few other people this summer, and I'm done with the majority of it, except for one feature. Part of what I've been doing is making it really user friendly, so that the program will prompt the user for certain pieces of information, and therefore know what needs to be randomized. I have it set up to check every piece of user input to make sure it's a valid input, and give an error message/prompt the user again if it's not. The only thing I can't quite figure out is how to get it to check whether or not the file name for the .csv output is valid. Does anyone know if there is a way to get R to check if a string makes a valid windows file name? Thanks!

Upvotes: 2

Views: 3571

Answers (2)

giocomai
giocomai

Reputation: 3528

I'd suggest that the easiest way to make sure a filename is valid is to use fs::path_sanitize().

It removes control characters, reserved characters, and Windows-reserved filenames, truncating the string at 255 bytes in length.

Upvotes: 3

Richie Cotton
Richie Cotton

Reputation: 121127

These characters aren't allowed: /\:*?"<>|. So warn the user if it contains any of those.

Some other names are also disallowed: COM, AUX, NUL, COM1 to COM9, LPT1 to LPT9.

You probably want to check that the filename is valid using a regular expression. See this other answer for a Java example that should take minimal tweaking to work in R. https://stackoverflow.com/a/6804755/134830

You may also want to check the filename length (260 characters for maximum portability, though longer names are allowed on some systems).

Finally, in R, if you try to create a file in a directory that doesn't exist, it will still fail, so you need to split the name up into the filename and directory name (using basename and dirname) and try to create the directory first, if necessary.

That said, David Heffernan gives good advice in his comment to let Windows do the wok in deciding whether or not it can create the file: you don't want to erroneously tell the user that a filename is invalid.

You want something a little like this:

nice_file_create <- function(filename)
{
  directory_name <- dirname(filename)
  if(!file.exists(directory_name))
  {
    ok <- dir.create(directory_name)
    if(!ok)
    {
      warning("The directory of that path could not be created.")
      return(invisible())
    }
  }
  tryCatch(
    file.create(filename),
    error = function(e)
    {
      warning("The file could not be created.")
    }
  )
}

But test it thoroughly first! There are all sorts of edge cases where things can fall over: try UNC network path names, "~", and paths with "." and ".." in them.

Upvotes: 2

Related Questions