Edit

Share via


How to stream XML fragments from an XmlReader (LINQ to XML)

When you have to process large XML files, it might not be feasible to load the whole XML tree into memory. This article shows how to stream fragments using an XmlReader in C# and Visual Basic.

One of the most effective ways to use an XmlReader to read XElement objects is to write your own custom axis method. An axis method typically returns a collection such as IEnumerable<T> of XElement, as shown in the example in this article. In the custom axis method, after you create the XML fragment by calling the ReadFrom method, return the collection using yield return. This provides deferred execution semantics to your custom axis method.

When you create an XML tree from an XmlReader object, the XmlReader must be positioned on an element. The ReadFrom method doesn't return until it has read the close tag of the element.

If you want to create a partial tree, you can instantiate an XmlReader, position the reader on the node that you want to convert to an XElement tree, and then create the XElement object.

The article How to stream XML fragments with access to header information contains information on streaming a more complex document.

The article How to perform streaming transform of large XML documents contains an example of using LINQ to XML to transform extremely large XML documents while maintaining a small memory footprint.

Example: Create a custom axis method

This example creates a custom axis method. You can query it by using a LINQ query. The custom axis method StreamRootChildDoc can read a document that has a repeating Child element.

static IEnumerable<XElement> StreamRootChildDoc(StringReader stringReader)
{
    using (XmlReader reader = XmlReader.Create(stringReader))
    {
        reader.MoveToContent();
        // Parse the file and display each of the nodes.
        while (reader.Read())
        {
            switch (reader.NodeType)
            {
                case XmlNodeType.Element:
                    if (reader.Name == "Child") {
                        XElement el = XElement.ReadFrom(reader) as XElement;
                        if (el != null)
                            yield return el;
                    }
                    break;
            }
        }
    }
}

static void Main(string[] args)
{
    string markup = @"<Root>
      <Child Key=""01"">
        <GrandChild>aaa</GrandChild>
      </Child>
      <Child Key=""02"">
        <GrandChild>bbb</GrandChild>
      </Child>
      <Child Key=""03"">
        <GrandChild>ccc</GrandChild>
      </Child>
    </Root>";

    IEnumerable<string> grandChildData =
        from el in StreamRootChildDoc(new StringReader(markup))
        where (int)el.Attribute("Key") > 1
        select (string)el.Element("GrandChild");

    foreach (string str in grandChildData) {
        Console.WriteLine(str);
    }
}
Module Module1
    Sub Main()
        Dim markup = "<Root>" &
                     "  <Child Key=""01"">" &
                     "    <GrandChild>aaa</GrandChild>" &
                     "  </Child>" &
                     "  <Child Key=""02"">" &
                     "    <GrandChild>bbb</GrandChild>" &
                     "  </Child>" &
                     "  <Child Key=""03"">" &
                     "    <GrandChild>ccc</GrandChild>" &
                     "  </Child>" &
                     "</Root>"

        Dim grandChildData =
             From el In New StreamRootChildDoc(New IO.StringReader(markup))
             Where CInt(el.@Key) > 1
             Select el.<GrandChild>.Value

        For Each s In grandChildData
            Console.WriteLine(s)
        Next
    End Sub
End Module

Public Class StreamRootChildDoc
    Implements IEnumerable(Of XElement)

    Private _stringReader As IO.StringReader

    Public Sub New(ByVal stringReader As IO.StringReader)
        _stringReader = stringReader
    End Sub

    Public Function GetEnumerator() As IEnumerator(Of XElement) Implements IEnumerable(Of XElement).GetEnumerator
        Return New StreamChildEnumerator(_stringReader)
    End Function

    Public Function GetEnumerator1() As IEnumerator Implements IEnumerable.GetEnumerator
        Return Me.GetEnumerator()
    End Function
End Class

Public Class StreamChildEnumerator
    Implements IEnumerator(Of XElement)

    Private _current As XElement
    Private _reader As Xml.XmlReader
    Private _stringReader As IO.StringReader

    Public Sub New(ByVal stringReader As IO.StringReader)
        _stringReader = stringReader
        _reader = Xml.XmlReader.Create(_stringReader)
        _reader.MoveToContent()
    End Sub

    Public ReadOnly Property Current As XElement Implements IEnumerator(Of XElement).Current
        Get
            Return _current
        End Get
    End Property

    Public ReadOnly Property Current1 As Object Implements IEnumerator.Current
        Get
            Return Me.Current
        End Get
    End Property

    Public Function MoveNext() As Boolean Implements IEnumerator.MoveNext
        While _reader.Read()
            Select Case _reader.NodeType
                Case Xml.XmlNodeType.Element
                    Dim el = TryCast(XElement.ReadFrom(_reader), XElement)
                    If el IsNot Nothing Then
                        _current = el
                        Return True
                    End If
            End Select
        End While

        Return False
    End Function

    Public Sub Reset() Implements IEnumerator.Reset
        _reader = Xml.XmlReader.Create(_stringReader)
        _reader.MoveToContent()
    End Sub

#Region "IDisposable Support"

    Private disposedValue As Boolean ' To detect redundant calls

    ' IDisposable
    Protected Overridable Sub Dispose(ByVal disposing As Boolean)
        If Not Me.disposedValue Then
            If disposing Then
                _reader.Close()
            End If
        End If
        Me.disposedValue = True
    End Sub

    Public Sub Dispose() Implements IDisposable.Dispose
        Dispose(True)
        GC.SuppressFinalize(Me)
    End Sub
#End Region

End Class

This example produces the following output:

bbb
ccc

The technique used in this example maintains a small memory footprint even for millions of Child elements.

See also