Path.Combine for URLs?

Question

Path.Combine is handy, but is there a similar function in the .NET framework for URLs?

I'm looking for syntax like this:

Url.Combine("http://MyUrl.com/", "/Images/Image.jpg")

which would return:

"http://MyUrl.com/Images/Image.jpg"

Ryan_43029 · Answer

You use Uri.TryCreate( ... ) :

Uri result = null;

if (Uri.TryCreate(new Uri("http://msdn.microsoft.com/en-us/library/"), "/en-us/library/system.uri.trycreate.aspx", out result))
{
    Console.WriteLine(result);
}

Will return:

http://msdn.microsoft.com/en-us/library/system.uri.trycreate.aspx

jayrdub_45767 · Answer

Path.Combine("Http://MyUrl.com/", "/Images/Image.jpg").Replace("\", "/")

mtazva_67538 · Answer

Witty example, Ryan, to end with a link to the function. Well done.

One recommendation Brian: if you wrap this code in a function, you may want to use a UriBuilder to wrap the base URL prior to the TryCreate call.

Otherwise, the base URL MUST include the scheme (where the UriBuilder will assume http://). Just a thought:

public string CombineUrl(string baseUrl, string relativeUrl) {
    UriBuilder baseUri = new UriBuilder(baseUrl);
    Uri newUri;

    if (Uri.TryCreate(baseUri.Uri, relativeUrl, out newUri))
        return newUri.ToString();
    else
        throw new ArgumentException("Unable to combine specified url values");
}

user_2136590 · Answer

As found in other answers, either new Uri() or TryCreate() can do the tick. However, the base Uri has to end with / and the relative has to NOT begin with /; otherwise it will remove the trailing part of the base Url

I think this is best done as an extension method, i.e.

public static Uri Append(this Uri uri, string relativePath)
{
    var baseUri = uri.AbsoluteUri.EndsWith('/') ? uri : new Uri(uri.AbsoluteUri + '/');
    var relative = relativePath.StartsWith('/') ? relativePath.Substring(1) : relativePath;
    return new Uri(baseUri, relative);
}

and to use it:

var baseUri = new Uri("http://test.com/test/");
var combinedUri =  baseUri.Append("/Do/Something");

In terms of performance, this consumes more resources than it needs, because of the Uri class which does a lot of parsing and validation; a very rough profiling (Debug) did a million operations in about 2 seconds. This will work for most scenarios, however to be more efficient, it's better to manipulate everything as strings, this takes 125 milliseconds for 1 million operations. I.e.

public static string Append(this Uri uri, string relativePath)
{
    //avoid the use of Uri as it's not needed, and adds a bit of overhead.
    var absoluteUri = uri.AbsoluteUri; //a calculated property, better cache it
    var baseUri = absoluteUri.EndsWith('/') ? absoluteUri : absoluteUri + '/';
    var relative = relativePath.StartsWith('/') ? relativePath.Substring(1) : relativePath;
    return baseUri + relative;
}

And if you still want to return a URI, it takes around 600 milliseconds for 1 million operations.

public static Uri AppendUri(this Uri uri, string relativePath)
{
    //avoid the use of Uri as it's not needed, and adds a bit of overhead.
    var absoluteUri = uri.AbsoluteUri; //a calculated property, better cache it
    var baseUri = absoluteUri.EndsWith('/') ? absoluteUri : absoluteUri + '/';
    var relative = relativePath.StartsWith('/') ? relativePath.Substring(1) : relativePath;
    return new Uri(baseUri + relative);
}

I hope this helps.

user_9918730 · Answer

I think this should give you more flexibility as you can deal with as many path segments as you want:

public static string UrlCombine(this string baseUrl, params string[] segments)
=> string.Join("/", new[] { baseUrl.TrimEnd('/') }.Concat(segments.Select(s => s.Trim('/'))));

user_13235578 · Answer

If you don't want to have a dependency like Flurl, you can use its source code:

    /// 
    /// Basically a Path.Combine for URLs. Ensures exactly one '/' separates each segment,
    /// and exactly on '&' separates each query parameter.
    /// URL-encodes illegal characters but not reserved characters.
    /// 
    /// URL parts to combine.
    public static string Combine(params string[] parts) {
        if (parts == null)
            throw new ArgumentNullException(nameof(parts));

        string result = "";
        bool inQuery = false, inFragment = false;

        string CombineEnsureSingleSeparator(string a, string b, char separator) {
            if (string.IsNullOrEmpty(a)) return b;
            if (string.IsNullOrEmpty(b)) return a;
            return a.TrimEnd(separator) + separator + b.TrimStart(separator);
        }

        foreach (var part in parts) {
            if (string.IsNullOrEmpty(part))
                continue;

            if (result.EndsWith("?") || part.StartsWith("?"))
                result = CombineEnsureSingleSeparator(result, part, '?');
            else if (result.EndsWith("#") || part.StartsWith("#"))
                result = CombineEnsureSingleSeparator(result, part, '#');
            else if (inFragment)
                result += part;
            else if (inQuery)
                result = CombineEnsureSingleSeparator(result, part, '&');
            else
                result = CombineEnsureSingleSeparator(result, part, '/');

            if (part.Contains("#")) {
                inQuery = false;
                inFragment = true;
            }
            else if (!inFragment && part.Contains("?")) {
                inQuery = true;
            }
        }
        return EncodeIllegalCharacters(result);
    }

    /// 
    /// URL-encodes characters in a string that are neither reserved nor unreserved. Avoids encoding reserved characters such as '/' and '?'. Avoids encoding '%' if it begins a %-hex-hex sequence (i.e. avoids double-encoding).
    /// 
    /// The string to encode.
    /// If true, spaces will be encoded as + signs. Otherwise, they'll be encoded as %20.
    /// The encoded URL.
    public static string EncodeIllegalCharacters(string s, bool encodeSpaceAsPlus = false) {
        if (string.IsNullOrEmpty(s))
            return s;

        if (encodeSpaceAsPlus)
            s = s.Replace(" ", "+");

        // Uri.EscapeUriString mostly does what we want - encodes illegal characters only - but it has a quirk
        // in that % isn't illegal if it's the start of a %-encoded sequence https://stackoverflow.com/a/47636037/62600

        // no % characters, so avoid the regex overhead
        if (!s.Contains("%"))
            return Uri.EscapeUriString(s);

        // pick out all %-hex-hex matches and avoid double-encoding 
        return Regex.Replace(s, "(.*?)((%[0-9A-Fa-f]{2})|$)", c => {
            var a = c.Groups[1].Value; // group 1 is a sequence with no %-encoding - encode illegal characters
            var b = c.Groups[2].Value; // group 2 is a valid 3-character %-encoded sequence - leave it alone!
            return Uri.EscapeUriString(a) + b;
        });
    }

user_5817690 · Answer

So I have another approach, similar to everyone who used UriBuilder.

I did not want to split my BaseUrl (which can contain a part of the path - e.g. http://mybaseurl.com/dev/) as javajavajavajavajava did.

The following snippet shows the code + Tests.

Beware: This solution lowercases the host and appends a port. If this is not desired, one can write a string representation by e.g. leveraging the Uri Property of UriBuilder.

  public class Tests
  {
         public static string CombineUrl (string baseUrl, string path)
         {
           var uriBuilder = new UriBuilder (baseUrl);
           uriBuilder.Path = Path.Combine (uriBuilder.Path, path);
           return uriBuilder.ToString();
         }

         [TestCase("http://MyUrl.com/", "/Images/Image.jpg", "http://myurl.com:80/Images/Image.jpg")]
         [TestCase("http://MyUrl.com/basePath", "/Images/Image.jpg", "http://myurl.com:80/Images/Image.jpg")]
         [TestCase("http://MyUrl.com/basePath", "Images/Image.jpg", "http://myurl.com:80/basePath/Images/Image.jpg")]
         [TestCase("http://MyUrl.com/basePath/", "Images/Image.jpg", "http://myurl.com:80/basePath/Images/Image.jpg")]
         public void Test1 (string baseUrl, string path, string expected)
         {
           var result = CombineUrl (baseUrl, path);

           Assert.That (result, Is.EqualTo (expected));
         }
  }

Tested with .NET Core 2.1 on Windows 10.

Why does this work?

Even though Path.Combine will return Backslashes (on Windows atleast), the UriBuilder handles this case in the Setter of Path.

Taken from https://github.com/dotnet/corefx/blob/master/src/System.Private.Uri/src/System/UriBuilder.cs (mind the call to string.Replace)

[AllowNull]
public string Path
{
      get
      {
          return _path;
      }
      set
      {
          if ((value == null) || (value.Length == 0))
          {
              value = "/";
          }
          _path = Uri.InternalEscapeString(value.Replace('\', '/'));
          _changed = true;
      }
 }

Is this the best approach?

Certainly this solution is pretty self describing (at least in my opinion). But you are relying on undocumented (at least I found nothing with a quick google search) "feature" from the .NET API. This may change with a future release so please cover the Method with Tests.

There are tests in https://github.com/dotnet/corefx/blob/master/src/System.Private.Uri/tests/FunctionalTests/UriBuilderTests.cs (Path_Get_Set) which check, if the \ is correctly transformed.

Side Note: One could also work with the UriBuilder.Uri property directly, if the uri will be used for a System.Uri ctor.

user_251267 · Answer

I have an allocation-free string creation version that I've been using with great success.

NOTE:

For the first string: it trims the separator using TrimEnd(separator) - so only from the end of the string.
For the remainders: it trims the separator using Trim(separator) - so both start and end of paths
It does not append a trailing slash/separator. Though a simple modification can be done to add this ability.

Hope you find this useful!

/// 
/// This implements an allocation-free string creation to construct the path.
/// This uses 3.5x LESS memory and is 2x faster than some alternate methods (StringBuilder, interpolation, string.Concat, etc.).
/// 
/// 
/// 
/// 
public static string ConcatPath(this string str, params string[] paths)
{
    const char separator = '/';
    if (str == null) throw new ArgumentNullException(nameof(str));

    var list = new List>();
    var first = str.AsMemory().TrimEnd(separator);

    // get length for intial string after it's trimmed
    var length = first.Length;
    list.Add(first);

    foreach (var path in paths)
    {
        var newPath = path.AsMemory().Trim(separator);
        length += newPath.Length + 1;
        list.Add(newPath);
    }

    var newString = string.Create(length, list, (chars, state) =>
    {
        // NOTE: We don't access the 'list' variable in this delegate since 
        // it would cause a closure and allocation. Instead we access the state parameter.

        // track our position within the string data we are populating
        var position = 0;

        // copy the first string data to index 0 of the Span
        state[0].Span.CopyTo(chars);

        // update the position to the new length
        position += state[0].Span.Length;

        // start at index 1 when slicing
        for (var i = 1; i < state.Count; i++)
        {
            // add a separator in the current position and increment position by 1
            chars[position++] = separator;

            // copy each path string to a slice at current position
            state[i].Span.CopyTo(chars.Slice(position));

            // update the position to the new length
            position += state[i].Length;
        }
    });
    return newString;
}

with Benchmark DotNet output:

|                Method |     Mean |    Error |   StdDev |   Median | Ratio | RatioSD |  Gen 0 | Allocated |
|---------------------- |---------:|---------:|---------:|---------:|------:|--------:|-------:|----------:|
| ConcatPathWithBuilder | 404.1 ns | 27.35 ns | 78.48 ns | 380.3 ns |  1.00 |    0.00 | 0.3347 |   1,400 B |
|            ConcatPath | 187.2 ns |  5.93 ns | 16.44 ns | 183.2 ns |  0.48 |    0.10 | 0.0956 |     400 B |

user_690656 · Answer

Why not just use the following.

System.IO.Path.Combine(rootUrl, subPath).Replace(@"\", "/")

user_2111229 · Answer

I found that the Uri constructor flips '\' into '/'. So you can also use Path.Combine, with the Uri constructor.

 Uri baseUri = new Uri("http://MyUrl.com");
 string path = Path.Combine("Images", "Image.jpg");
 Uri myUri = new Uri(baseUri, path);

user_2574087 · Answer

For what it's worth, here a couple of extension methods. The first one will combine paths and the second one adds parameters to the URL.

    public static string CombineUrl(this string root, string path, params string[] paths)
    {
        if (string.IsNullOrWhiteSpace(path))
        {
            return root;
        }

        Uri baseUri = new Uri(root);
        Uri combinedPaths = new Uri(baseUri, path);

        foreach (string extendedPath in paths)
        {
           combinedPaths = new Uri(combinedPaths, extendedPath);
        }

        return combinedPaths.AbsoluteUri;
    }

    public static string AddUrlParams(this string url, Dictionary parameters)
    {
        if (parameters == null || !parameters.Keys.Any())
        {
            return url;
        }

        var tempUrl = new StringBuilder($"{url}?");
        int count = 0;

        foreach (KeyValuePair parameter in parameters)
        {
            if (count > 0)
            {
                tempUrl.Append("&");
            }

            tempUrl.Append($"{WebUtility.UrlEncode(parameter.Key)}={WebUtility.UrlEncode(parameter.Value)}");
            count++;
        }

        return tempUrl.ToString();
    }

user_197591 · Answer

If you don't want to add a third-party dependency such as Flurl or create a custom extension method, in ASP.NET Core (also available in Microsoft.Owin), you can use PathString which is intended for the purpose of building up URI paths. You can then create your full URI using a combination of this, Uri and UriBuilder.

In this case, it would be:

new Uri(new UriBuilder("http", "MyUrl.com").Uri, new PathString("/Images").Add("/Image.jpg").ToString())

This gives you all the constituent parts without having to specify the separators in the base URL. Unfortunately, PathString requires that / is prepended to each string otherwise it in fact throws an ArgumentException! But at least you can build up your URI deterministically in a way that is easily unit-testable.

user_9822528 · Answer

For anyone who is looking for a one-liner and simply wants to join parts of a path without creating a new method or referencing a new library or construct a URI value and convert that to a string, then...

string urlToImage = String.Join("/", "websiteUrl", "folder1", "folder2", "folder3", "item");

It's pretty basic, but I don't see what more you need. If you're afraid of doubled '/' then you can simply do a .Replace("//", "/") afterward. If you're afraid of replacing the doubled '//' in 'https://', then instead do one join, replace the doubled '/', then join the website url (however I'm pretty sure most browsers will automatically convert anything with 'https:' in the front of it to read in the correct format). This would look like:

string urlToImage = String.Join("/","websiteUrl", String.Join("/", "folder1", "folder2", "folder3", "item").Replace("//","/"));

There are plenty of answers here that will handle all the above, but in my case, I only needed it once in one location and won't need to heavily rely on it. Also, it's really easy to see what is going on here.

See: https://docs.microsoft.com/en-us/dotnet/api/system.string.join?view=netframework-4.8

user_869033 · Answer

A simple one liner:

public static string Combine(this string uri1, string uri2) => $"{uri1.TrimEnd('/')}/{uri2.TrimStart('/')}";

Inspired by @Matt Sharpe's answer.

user_4069409 · Answer

Recently Combine method was added to Energy.Core package, so you might want to use it to join URL parts.

    string url;
    url = Energy.Base.Url.Combine("https://www.youtube.com", "watch?v=NHCgbs3TcYg");
    Console.WriteLine(url);
    url = Energy.Base.Url.Combine("https://www.youtube.com", "watch?v=NHCgbs3TcYg", "t=150");
    Console.WriteLine(url);

Additionally it will recognize parameter part, so it will work as you might expect (joining path with slash and parameters with ampersand).

https://www.youtube.com/watch?v=NHCgbs3TcYg

https://www.youtube.com/watch?v=NHCgbs3TcYg&t=150

Documentation for Energy.Base.Url class

Package on NuGet gallery

Code sample

user_2368285 · Answer

// Read all above samples and as result created my self:

static string UrlCombine(params string[] items)
{
    if (items?.Any() != true)
    {
        return string.Empty;
    }

    return string.Join("/", items.Where(u => !string.IsNullOrWhiteSpace(u)).Select(u => u.Trim('/', '\')));
}

// usage

UrlCombine("https://microsoft.com","en-us")

Path.Combine for URLs?

42 Answers