When we tried to serialize an object to XML we noticed that Nullable types are special. A class is not serialized when null, but a Nullable type is a struct. See the code blow and notice the highlighted line in the output.

class Program
{
    static void Main(string[] args)
    {
        var obj = new MyType();
        var serializer = new XmlSerializer(typeof(MyType));
        serializer.Serialize(Console.Out, obj);
        Console.ReadLine();
    }
}

[Serializable]
public class MyType
{
    public string        MyStringProperty      { get; set; }
    public int           MyIntProperty         { get; set; }
    public Nullable<int> MyNullableIntProperty { get; set; }
}

Unwanted output:

<?xml version="1.0" encoding="ibm850"?>
<MyType>
  <MyIntProperty>0</MyIntProperty>
  <MyNullableIntProperty xsi:nil="true" />
</MyType>

The dotnet framework has a trick up its sleeve. Use the ShouldSerialize{PropertyName} function to manipulate the way a property is serialized. The corrected class is below. Using the same program the output now was without the (ugly) line of the nullable property.

[Serializable]
public class MyType
{
    public string        MyStringProperty      { get; set; }
    public int           MyIntProperty         { get; set; }
    public Nullable<int> MyNullableIntProperty { get; set; }
    // do not serialize MyNullableIntProperty if it has not been set
    public bool ShouldSerializeMyNullableIntProperty() 
    { 
        return MyNullableIntProperty.HasValue; 
    }
}

Correct output:

<?xml version="1.0" encoding="ibm850"?>
<MyType>
  <MyIntProperty>0</MyIntProperty>
</MyType>