CODEWITHSUNDEEP
c#.netfile
Matt Mitchell
Matt Mitchell

Reputation: 41861

How do you extract the directory from a path including wildcards?

Given paths like this:

C:\Temp\SomeDirectory\*.xml

I'd like to be able to distinguish the *.xml from C:\Temp\SomeDirectory

However, I do not want a path to a directory that does not have a trailing slash to return its parent directory.

This means I want the following behaviour:

// Wildcard paths return directory
C:\Temp\SomeDirectory\*.csv -> C:\Temp\SomeDirectory

// Trailing slash paths return full path
C:\Temp\SomeDirectory\ -> C:\Temp\SomeDirectory     

// Non-trailing slash paths to a directory return full path
C:\Temp\SomeDirectory -> C:\Temp\SomeDirectory

// Paths to a file return the directory
C:\Temp\SomeDirectory\SomeFileThatExists.csv -> C:\Temp\SomeDirectory

// Paths to a file without an extension (that exists) return the directory
C:\Temp\SomeDirectory\SomeFileThatExistsWithNoExt -> C:\Temp\SomeDirectory

// Paths to a non-existent path without a trailing slash are standard
// Either always clip the trailing part, or always leave it in
// (Can live with this one being C:\Temp\SomeDirectory)
C:\Temp\SomeDirectory\NonExistentObject -> C:\Temp\SomeDirectory\NonExistentObject 

// Paths to a non-existent path with a trailing slash return the full path
C:\Temp\SomeDirectory\NonExistentObject\ -> C:\Temp\SomeDirectory\NonExistentObject

// Paths to a non-existent path with a file extension return the directory
C:\Temp\SomeDirectory\NonExistentFile.Ext -> C:\Temp\SomeDirectory

(I am not fussed if the return value has a trailing slash or not, although the method I put below consistently does not return a trailing slash)

My current code is something like this, and handles these cases:

public string GetDirectory(string path)
{
    try
    {
        var f = new FileInfo(path); // Throws if invalid path, e.g. wildcards

        // Existent directory
        if (Directory.Exists(path))
        {
            // Full path must be a directory, so return full path
            // Ensure to add a trailing slash, as if it's missing it will return parent directory
            return Path.GetDirectoryName(path + '/');
        }

        // Non-existent directory (or ambiguous path without an extension or trailing slash)
        if (!System.IO.File.Exists(path) && String.IsNullOrEmpty(Path.GetExtension(path)))
        {
            // Path is to a non-existent file (without an extension) or to a non-existent directory.
            // As the path does not exist we will standardise and treat it as a directory.
            return Path.GetDirectoryName(path + '/');
        }

        // Path is to a file, return directory
        return Path.GetDirectoryName(path);
    }
    catch (ArgumentException)
    {
        // For wildcards/invalid paths, return the directory
        // This maps C:\Dir\*.csv to C:\Dir
        // Also maps C:\Dir&*A*#$!@& to C:\
        return Path.GetDirectoryName(path);
    }
}

Is there a better way to achieve this behaviour, or my ultimate goal of being able to get the "directory" from a path that may include wildcards?

Upvotes: 2

Views: 1301

Answers (2)

Matthew Strawbridge
Matthew Strawbridge

Reputation: 20640

Do you have any control over the generation of the list of paths? If you could ensure that all paths to directories end with a trailing slash (which I believe is the convention) then the simple Path.GetDirectoryName(path) will work for all these cases.

Upvotes: 2

Brandon Moore
Brandon Moore

Reputation: 8790

The problem I believe is that the Path methods are just string manipulation functions. I don't believe they actually go out and find whether you're looking at a directory or a file without an extension.

You will need to use the Directory or File class in conjunction to find out which it is and then manually change it accordingly.

Upvotes: 2

Related Questions