C# – the correct way to serialize nullable types

cixmlserializablenullableserialization

I'm implementing IXMLSerializable in one of my classes. It contains a few numeric properties that are nullable (int? double? etc..)

What is the correct way to serialize/serialize these through IXMLSerializable? Here's what I'm doing now, which works, but obviously does not seem like the correct way to do it.

void IXmlSerializable.WriteXml(XmlWriter writer)
{
    ...

    if (this._PropName == null)
    {
        writer.WriteElementString("PropName", "NULL");
    }
    else
    {
        writer.WriteElementString("PropName", this._PropName.ToString());
    }
    ...
}

void IXmlSerializable.ReadXml(XmlReader reader)    
{
    string tempStr;
    ...

    reader.ReadStartElement("PropName"); 

    if (tempStr != "NULL")
    {
        this._PropName = double.Parse(tempStr);
    }
    else 
    {
        this._PropName = null;
    }
    ...
}

Update:
It was asked that I give a little background as to why I am implementing IXmlSerializable. I'm working on a program for architectural design, where I need a class that represents a collection of Floors. Each Floor has properties such as Floor.Area area, Floor.Height etc. The elevation of the floor, however, is defined by the sum of the floor heights below it. So whenever a Floor.Height property changes or the FloorCollection is modified the elevations of the Floors are re-cacluated.

My FloorCollection class, which I need to serialize, inherits from BindingList. If I try to serialize this class directly it will serialize the collection of floors, but not any properties or fields in the class. See my previous post on this.

Now I'm trying to add the ability to restrict the maximum height, maximum top elevation, and minimum bottom elevation of the building floors in the collection. So I'm using nullable doubles to represent these restrictions, where a null value means unrestricted. The elevation properties can be positive, negative, or zero. So there needs to be an alternate state, null, that identifies when there is no restriction.

Now that I think about it might be easier overall to just have a separate Boolean value that identifies if there is a elevation/height restriction, and then a regular double property that identifies what the restriction is if it is enabled.

Best Answer

You want to always write the XML for the property, but if the property value is null you want to include an xsi:nil="true" attribute.

void IXmlSerializable.WriteXml(XmlWriter writer)
{
    ...

    if (this._PropName == null)
    {
        writer.WriteStartElement("PropName");
        writer.WriteAttributeString("xsi", "nil", "http://www.w3.org/2001/XMLSchema-instance", "true");
        writer.WriteEndElement();
    }
    else
    {
        writer.WriteElementString("PropName", this._PropName.ToString());
    }
    ...
}

You are also probably going to want to write an xsi:type="xsd:datatype" attribute, where xsd is the http://www.w3.org/2001/XMLSchema namespace. This will allow you to read the data type back in during deserialization to know if (and how) to convert the value.

This is the style that Microsoft tends to use in their examples.

It appears that the guidance in this area may have changed, as StyleCop now enforces the use of the C# specific aliases.

C# – What does the [Flags] Enum Attribute mean in C#

The [Flags] attribute should be used whenever the enumerable represents a collection of possible values, rather than a single value. Such collections are often used with bitwise operators, for example:

var allowedColors = MyColor.Red | MyColor.Green | MyColor.Blue;

Note that the [Flags] attribute doesn't enable this by itself - all it does is allow a nice representation by the .ToString() method:

enum Suits { Spades = 1, Clubs = 2, Diamonds = 4, Hearts = 8 }
[Flags] enum SuitsFlags { Spades = 1, Clubs = 2, Diamonds = 4, Hearts = 8 }

...

var str1 = (Suits.Spades | Suits.Diamonds).ToString();
           // "5"
var str2 = (SuitsFlags.Spades | SuitsFlags.Diamonds).ToString();
           // "Spades, Diamonds"

It is also important to note that [Flags] does not automatically make the enum values powers of two. If you omit the numeric values, the enum will not work as one might expect in bitwise operations, because by default the values start with 0 and increment.

Incorrect declaration:

[Flags]
public enum MyColors
{
    Yellow,  // 0
    Green,   // 1
    Red,     // 2
    Blue     // 3
}

The values, if declared this way, will be Yellow = 0, Green = 1, Red = 2, Blue = 3. This will render it useless as flags.

Here's an example of a correct declaration:

[Flags]
public enum MyColors
{
    Yellow = 1,
    Green = 2,
    Red = 4,
    Blue = 8
}

To retrieve the distinct values in your property, one can do this:

if (myProperties.AllowedColors.HasFlag(MyColor.Yellow))
{
    // Yellow is allowed...
}

or prior to .NET 4:

if((myProperties.AllowedColors & MyColor.Yellow) == MyColor.Yellow)
{
    // Yellow is allowed...
}

if((myProperties.AllowedColors & MyColor.Green) == MyColor.Green)
{
    // Green is allowed...
}

Under the covers

This works because you used powers of two in your enumeration. Under the covers, your enumeration values look like this in binary ones and zeros:

 Yellow: 00000001
 Green:  00000010
 Red:    00000100
 Blue:   00001000

Similarly, after you've set your property AllowedColors to Red, Green and Blue using the binary bitwise OR | operator, AllowedColors looks like this:

myProperties.AllowedColors: 00001110

So when you retrieve the value you are actually performing bitwise AND & on the values:

myProperties.AllowedColors: 00001110
             MyColor.Green: 00000010
             -----------------------
                            00000010 // Hey, this is the same as MyColor.Green!

The None = 0 value

And regarding the use of 0 in your enumeration, quoting from MSDN:

[Flags]
public enum MyColors
{
    None = 0,
    ....
}

Use None as the name of the flag enumerated constant whose value is zero. You cannot use the None enumerated constant in a bitwise AND operation to test for a flag because the result is always zero. However, you can perform a logical, not a bitwise, comparison between the numeric value and the None enumerated constant to determine whether any bits in the numeric value are set.

You can find more info about the flags attribute and its usage at msdn and designing flags at msdn

Best Answer

Related Solutions

C# – the difference between String and string in C#

This is the style that Microsoft tends to use in their examples.

C# – What does the [Flags] Enum Attribute mean in C#

Related Topic