In a Windows app I created I did let the users supply a XML file/value which it then converted to an in-memory Sqlite database. After that they can use "simple" SQL queries to extract the desired XML content. Such a class would be great for multiple purposes.

NB: XML is basically the same as a relational database, with related parent/child nodes.

Or, not convert it to a relational database, but flatten it like this:

<parent>
   <title>abc</title>
   <child>
      <name>xyz</name>
   </child>
   <child>
      <name>pqr</name>
   </child></parent>

... could be converted to a dataset like this:

parent_title | parent_child_name
abc | xyz
abc | pqr

@marco Well, it would work in the simple case you provided but there are a lot more cases where a solution would not so easy. What about attributes for example?

A special query language, like LINQ in C#, would be nice, but I always found the XML to SQLite solution working quite well, even with more complex XML structures. An attribute can basically be seen as a child of the element.

<parent>
   <title>abc</title>
   <child gender="male">
      <name>xyz</name>
   </child>
   <child gender="female">
      <name>pqr</name>
   </child>
</parent>

is effectively the same as:

<parent>
   <title>abc</title>
   <child>
      <gender>male</gender>
      <name>xyz</name>
   </child>
   <child>
      <gender>female</gender>
      <name>pqr</name>
   </child>
</parent>

and can be flatten like:

parent_title | parent_child_gender | parent_child_name
abc | male | xyz
abc | female | pqr

Of course there could be more XML parts like namespaces, but I just used those as prefixes. At least that's the way I handled it.

Again a query language would be better, but I personally found this a "good enough" solution. Good enough for me ;)