Question

I have an XSD and I have to generate an XML document to send to the customers of the company I work with. The documents I send will be validated against this XSD schema.

What is the best way to create a XML document conforming to a XSD Schema? I mean, I’m searching for best practices and the like. I’m new to this and while “Googling” around here and there, I found people using XmlTextWriter, DataSet.WriteXml, and others.

1. DataSet.WriteXml seems to not work well for me. This is what I did:

   var ds = new DataSet();
   ds.ReadXmlSchema(schemaFile);
   ds.Tables["TableName"].Rows.Add("", "", 78, true, DateTime.Now);
   ...
   ds.WriteXml("C:\\xml.xml");
   

   I found it generates a node with NewDataSet, and the nodes are not in the proper order.

2. XmlTextWriter, I find it a bit long to do… but I will if there is no other choice.

What do you think is the best way to do this? Are there other approaches to do it?
I would put the schema here if it wasn’t so long, and if it were relevant to the question.

Answer 1

The mainstream practice in .NET is to use XML Serialization.

In your case, I would do this:

• run the xsd.exe too on .XSD to generate source code for classes (xsd /c)
• build your app that utilizes those generated classes. Keep in mind you can extend those classes via the “partial classes” technique
• in code, instantiate an XmlSerializer, and Serialize the class instances.

Example:

Given this schema:

<xs:schema elementFormDefault="qualified" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element name="Foo" nillable="true" type="Foo" />
  <xs:complexType name="Foo">
    <xs:sequence>
      <xs:element minOccurs="0" maxOccurs="1" name="Bar" type="xs:string" />
      <xs:element minOccurs="0" maxOccurs="1" name="Baz" type="UntypedArray" />
    </xs:sequence>
  </xs:complexType>


  <xs:complexType name="UntypedArray">
    <xs:choice minOccurs="1" maxOccurs="unbounded">
      <xs:element name="Type1" type="Type1"                 minOccurs="1" maxOccurs="1"/>
      <xs:any     namespace="##other" processContents="lax" minOccurs="1" maxOccurs="1"/>
    </xs:choice>
  </xs:complexType>


  <xs:complexType name="Type1" mixed="true">
    <xs:sequence>
      <xs:element minOccurs="0" maxOccurs="1" name="Child" type="xs:string" />
    </xs:sequence>
  </xs:complexType>
</xs:schema>

xsd.exe generates this source code:

[System.CodeDom.Compiler.GeneratedCodeAttribute("xsd", "2.0.50727.42")]
[System.SerializableAttribute()]
[System.Diagnostics.DebuggerStepThroughAttribute()]
[System.ComponentModel.DesignerCategoryAttribute("code")]
[System.Xml.Serialization.XmlRootAttribute(Namespace="", IsNullable=true)]
public partial class Foo {

    private string barField;

    private object[] bazField;

    /// <remarks/>
    public string Bar {
        get {
            return this.barField;
        }
        set {
            this.barField = value;
        }
    }

    /// <remarks/>
    [System.Xml.Serialization.XmlArrayItemAttribute("", typeof(System.Xml.XmlElement), IsNullable=false)]
    [System.Xml.Serialization.XmlArrayItemAttribute(typeof(Type1), IsNullable=false)]
    public object[] Baz {
        get {
            return this.bazField;
        }
        set {
            this.bazField = value;
        }
    }
}

/// <remarks/>
[System.CodeDom.Compiler.GeneratedCodeAttribute("xsd", "2.0.50727.42")]
[System.SerializableAttribute()]
[System.Diagnostics.DebuggerStepThroughAttribute()]
[System.ComponentModel.DesignerCategoryAttribute("code")]
public partial class Type1 {

    private string childField;

    private string[] textField;

    /// <remarks/>
    public string Child {
        get {
            return this.childField;
        }
        set {
            this.childField = value;
        }
    }

    /// <remarks/>
    [System.Xml.Serialization.XmlTextAttribute()]
    public string[] Text {
        get {
            return this.textField;
        }
        set {
            this.textField = value;
        }
    }
}

In your app you can instantiate a Foo and then serialize, like this:

    Foo foo = new Foo();
    // ...populate foo here...
    var builder = new System.Text.StringBuilder();
    XmlSerializer s = new XmlSerializer(typeof(Foo));
    using ( var writer = System.Xml.XmlWriter.Create(builder))
    {
        s.Serialize(writer, foo, ns);
    }
    string rawXml = builder.ToString();

This example serializes into a string. Of course you can serialize to other XmlWriters, you can write out to a file, to any arbitrary stream, and so on.

Normally I tweak the serialization to omit the XML declaration, omit the default xml namespaces, and so on. Like this:

    Foo foo = new Foo();
    // ...populate foo here...
    var builder = new System.Text.StringBuilder();
    var settings = new System.Xml.XmlWriterSettings { OmitXmlDeclaration = true, Indent= true };
    var ns = new XmlSerializerNamespaces();
    ns.Add("","");
    XmlSerializer s = new XmlSerializer(typeof(Foo));
    using ( var writer = System.Xml.XmlWriter.Create(builder, settings))
    {
        s.Serialize(writer, foo, ns);
    }
    string rawXml = builder.ToString();

You can also do the reverse – map from an XML document to an in-memory object graph – using the XmlSerializer. Use the Deserialize method.