C#: How would I filter out the unneeded namespaces in this scenario?

Question

This is more of a language-agnostic question than a C#-specific one, but since it deals with namespaces I thought I'd tag this as .NET-related.

Say you have a bunch of strings that represent namespaces:

[
    "System",
    "System.Windows",
    "System.Windows.Input",
    "System.Windows.Converters",
    "System.Windows.Markup",
    "System.Windows.Markup.Primitives",
    "System.IO",
    "System.IO.Packaging"
]

You want to filter them in such a way that any namespace that 'contains' another namespace in the collection is excluded. For example, since System contains System.Windows , it's excluded. Since System.IO is a parent of System.IO.Packaging , it's likewise excluded. This continues until you end up with this:

[
    "System.Windows.Input",
    "System.Windows.Converters",
    "System.Windows.Markup.Primitives",
    "System.IO.Packaging"
]

What would be an efficient way to filter a list this way in C#? I'm looking for a method that will look something like this:

public IEnumerable<string> FilterNamespaces(IEnumerable<string> namespaces) {}

Any help would be appreciated, thanks!

---

EDIT: Here is what ultimately worked for me:

public IEnumerable<string> FilterNamespaces(IEnumerable<string> namespaces) =>
    namespaces.Where(current => !namespaces.Any(n => n.StartsWith($"{current}.")));

Answer 1

Try following

 public static IEnumerable<string> FilterNamespaces(IEnumerable<string> namespaces)
 => namespaces
  .Where(ns => namespaces
    .Where(n => n != ns)
      .All(n => !Regex.IsMatch(n, $@"{Regex.Escape(ns)}[\.\n]")))
  .Distinct();   

Answer 2

You can build a tree where the root is the System namespace and every edge will represent a parent-child relation between different namespaces. Then you can find all of the leafs of the tree with a traversing algorithm which finds the rank of every node(nodes with rank=1 are leafs-they do not have children).