How to use Path.GetInvalidPathChars to find a file name inside a larger string?

Question

My goal is to find a file name ("MyFile.txt") inside a larger string. I.e.:

Some text before MyFile.txt some other text after

Currently I'm successfully using a Regular Expression with a character class of something like the following (simplified):

[\w\.\-]

This works fine, until the file contains other characters that are outside the \w group, e.g. an em dash: "My—File.txt".

My approach:

The method Path.GetInvalidPathChars returns an array of invalid characters. I've tried to use this method. Unfortunately, I found no way of "converting" this to be useful inside a Regular Expression.

I'm aware of

The SO posting "How to remove illegal characters from path and filenames?"
The concept of "Character class subtraction"

Still, I found no solution.

My question:

Is there any Regular Expression (or any other way) to find and extract a file name inside a larger string, based on the result of Path.GetInvalidPathChars?

This is a very difficult thing - there is no safe regex way to extract arbitrary file names from larger texts. — Wiktor Stribiżew, May 22 '17 at 09:37
You´d have to build a regex for searching strings that do *not* contain any of the invalid characters. But this could be *anything*, even files without extension exist - at least on Windows. — MakePeaceGreatAgain, May 22 '17 at 09:38
@jdweng I'm doing it for convenience, since a dash is a special character in `a-z`. — Uwe Keim, May 22 '17 at 09:39
You also have to limit what the names of the file can be - e.g. `.txt` files. Files don't need to have extensions. or can have spaces and "@" characters. or start with a dot. — Martin Ullrich, May 22 '17 at 09:39
If you _had_ to do it using re, how about https://regex101.com/r/dRWjqr/1? As most have already pointed out, this is a very difficult task, and would need a lot of assumptions on the dataset for it to work correctly — degant, May 22 '17 at 09:56
Ideally, I think I a feature to define my own negative character group (like `\W`) would be the solution. Unfortunately, this is not possible in .NET. — Uwe Keim, May 22 '17 at 10:20

MakePeaceGreatAgain · Accepted Answer · 2017-05-22T11:59:22.413

3

I won´t use a regex for this at all as it becomes incredibly complex and unreadable. In particular a filename could be nearly any string, including most special characters, numbers, spaces. Even worse there are even files without a dot to seperate an extension. So I´d suggest to simply do an Contains-check on all your invalid characters:

char[] invalidChars = Path.GetInvalidPathChars;
bool valid = !myString.Contains(x => invalidChars.Contains(x));

Extracting the candidates instead is even simpler. The idea is to split your large string on all invalid characters. This means everything in between the invalid characters is considered a file-name, e.g:

"myTest.extension" → "myTest.extension"
"myFile:anotherFile" → "myFile"; "anotherFile"
"myFile with space" → "myFile with space"
"a File with .-determined extension.dot" → "a File with .-determined extension.dot"

This is achieved by this code:

var fileNames = myText.Split(invalidChars);

EDIT: If you really want a regex you can build one dynamically from your invalid characters:

var pattern = String.Format("([^{0}]*)", new String(invalidCharacters));
var r = new Regex(pattern);

edited May 22 '17 at 11:59

answered May 22 '17 at 09:41

MakePeaceGreatAgain

35,491
6
60
111

Only part of my string is a file name, not the whole string. Plus: I do not want to check true/false, I want to get the string. – Uwe Keim May 22 '17 at 09:41
No -1 from me. The `Split` idea looks nice! – Uwe Keim May 22 '17 at 09:51
1

Instead `String.Join("", invalidCharacters.Select(x => x.ToString()))` I do prefer [`new string(invalidCharacters)`](https://msdn.microsoft.com/en-us/library/ttyxaek9(v=vs.110).aspx) . – Uwe Keim May 22 '17 at 10:58
1

`String.Join("", ` is now unnecessary, I guess. I've removed it in your code, hope this is OK. – Uwe Keim May 22 '17 at 11:50

Hossein Narimani Rad · Answer 2 · 2017-05-22T10:01:41.490

0

If your file name do not contains space and do contain extension, then this simple idea may help you

string line = "Some text before MyFile.txt some other text after";

//If you look for path:
//var array = Path.GetInvalidPathChars().ToList();

//If you look for file name
var array = Path.GetInvalidFileNameChars().ToList();
array.Add(' ');

var potentialFileNames = line.Split(array.ToArray(), StringSplitOptions.RemoveEmptyEntries)
                             .Where(i => i.Contains('.')).ToList();

 //potentialFileNames[0] = "MyFile.txt"

edited May 22 '17 at 10:01

answered May 22 '17 at 09:50

Hossein Narimani Rad

31,361
18
86
116

I convert it to add the space before using it in split – Hossein Narimani Rad May 22 '17 at 09:52

How to use Path.GetInvalidPathChars to find a file name inside a larger string?

2 Answers2