从目录路径提取子字符串的最佳方法

Question

I have below strings

str1 = "C:\Program Files (x86)\Server\Client\v1.2\Service\test.exe"
str1 = "C:\Program Files (x86)\Server\v1.2\Service\test.exe"
str1 = "C:\Program Files (x86)\Server\v3\Service\test.exe"
str1 = "C:\Program Files (x86)\Client\v3.6\Service\test.exe"

I need to extract string till V1.2 or V3 or V3.6 like below,

str1 = "C:\Program Files (x86)\Server\Client\"
str1 = "C:\Program Files (x86)\Server\"
str1 = "C:\Program Files (x86)\Server\"
str1 = "C:\Program Files (x86)\Client\"

What should be the best way, regex?

Answer 1

Regex.Split(str1, @"(?<=\\)(v|V)\d+(\.\d+)?\\")[0]

• (<?=\\\\) is look behind for \\ , making sure version is a directory name.
• (v|V) case-intensive
• \\d+ a number
• (\\.\\d+)? may not have minor version
• \\\\ ending path splitor.

Answer 2

If you don't want to deal with manual path parsing, you can use .NET to do the heavy lifting:

static readonly Regex VersionRegex = new Regex(@"^v\d+(\.\d+)?$", 
                                               RegexOptions.Compiled | RegexOptions.IgnoreCase);

public static string SkipToApplicationRoot(string path)
{
  var di = new DirectoryInfo(path);

  while (di.Parent != null)
  {
    // If the current directory is a version string, return the path to the parent directory
    if (VersionRegex.IsMatch(di.Name)) return di.Parent.FullName;

    di = di.Parent;
  }

  // No version string in the path
  return null;
}

In English, take the deepest folder in the path, verify if it's in the version string format, if not, repeat with the parent folder. If it is a version string, take the full path of the parent folder. If we reach the root folder, return null (no version string in path).

You can tweak VersionRegex to accept exactly the kinds of version strings you want to consider. Of course, the whole algorithm is only as reliable as your assumption that there's only one folder that fits the version string format between the full path and the root you're trying to find.

There's probably simpler code that fits your actual requirements. For the strings you've actually posted, simply going two directories up in the structure would also suffice. But as it is often said, if you can precisely formulate your requirements, you probably also have the solution already :) Given that you've given no reasoning behind what you're trying to do, it's quite possible this is an XY problem.

Answer 3

You can split by '\\' then use a regex to find the exact match.

Test the version regex ^v[0-9]*(\\.[0-9]{1,2})?$ here . I have assumed that your version will have upto two decimal places.

You can try like this:

---

static void Main(string[] args)
{
    string str1 = @"C:\Program Files (x86)\Server\Client\v1.2\Service\test.exe";
    string str2 = @"C:\Program Files (x86)\Server\v1.2\Service\test.exe";
    string str3 = @"C:\Program Files (x86)\Server\v3\Service\test.exe";
    string str4 = @"C:\Program Files (x86)\Client\v3.6\Service\test.exe";

    List<string> versions = FetchVersionPaths(str1, str2, str3, str4).ToList();

    versions.ForEach(Console.WriteLine);
}

public static IEnumerable<string> FetchVersionPaths(params string[] DirectoryPaths)
{
    foreach (string directoryPath in DirectoryPaths)
    {
        // Version regex to check upto two decimal places
        Regex versionRegex = new Regex(@"^v[0-9]*(\.[0-9]{1,2})?$");

        List<string> destructuredPaths = directoryPath.Split(Path.DirectorySeparatorChar).ToList();

        // Find the index upto version
        int versionSplitIndex = destructuredPaths.FindIndex(versionRegex.IsMatch);

        // Now get the range of the destructured paths and join them by '\'
        yield return string.Join(new string(Path.DirectorySeparatorChar, 1), destructuredPaths.GetRange(0, versionSplitIndex));
    }
}

---

Outputting:

C:\Program Files (x86)\Server\Client
C:\Program Files (x86)\Server
C:\Program Files (x86)\Server
C:\Program Files (x86)\Client