Can I get the class name in a static method or property

[steve]

If I have an object I know I can get the name of the class but can I get the class name in the following scenario :

public class ParentClass
{
public static string ClassName
{
get
{
return "ParentClas s"
}

}
}
public class ChildClass : ParentClass
{

}
So that ChildClass.Clas sName returns "ChildClass " and ParentClass.Cla ssName returns "ParentClas s"

It may seem a bit strange but I really want to use it to return a token based on the class name that can be used for reading data
from an xml file and that contains a list of these classes.

I would like to be able to base the read on just knowing the class name and not have to store static data in each class which could
be calculated from the class name.

So, I will be able to write

MyStoredData.Ge tData(ChildClas s.ClassName) and it will return all the records in the xml of type ChildClass and MyStoredData
encapsulates the reading of data and GetData is a static method
It is not important whether I get the class name from a method or a property.

Any ideas welcome. Hope it makes sense.

Steve

[Marc Gravell]

The cleanest answer I can suggest is generics - have GetData<T>(), so
you call GetData<ChildCl ass>(), then inside GetData<T>() you can use
typeof(T).Name, or access attribute metadata from typeof(T) such as
XmlTypeAttribut e.

Marc

[Jon Skeet [C# MVP]]

steve <s_******@yahoo .comwrote:

>
If I have an object I know I can get the name of the class
but can I get the class name in the following scenario :

public class ParentClass
{
public static string ClassName
{
get
{
return "ParentClas s"
}

}
}
public class ChildClass : ParentClass
{

}
So that ChildClass.Clas sName returns "ChildClass "
and ParentClass.Cla ssName returns "ParentClas s"

No, you can't do this. In fact, if you look at the generated code for
"ChildClass.Cla ssName" you'll see it's *actually* a call to
"ParentClass.Cl assName".

It may seem a bit strange but I really want to use it to return a
token based on the class name that can be used for reading data from
an xml file and that contains a list of these classes.

I would like to be able to base the read on just knowing the class
name and not have to store static data in each class which could be
calculated from the class name.

So, I will be able to write

MyStoredData.Ge tData(ChildClas s.ClassName) and it will return all the
records in the xml of type ChildClass and MyStoredData encapsulates
the reading of data and GetData is a static method

It is not important whether I get the class name from a method or a
property.

Any ideas welcome. Hope it makes sense.

Could you change to use generics instead?

MyStoredData.Ge tData<ChildClas s>(...)

?

--
Jon Skeet - <sk***@pobox.co m>
Web site: http://www.pobox.com/~skeet
Blog: http://www.msmvps.com/jon.skeet
C# in Depth: http://csharpindepth.com

[Peter Duniho]

On Tue, 02 Sep 2008 13:33:25 -0700, steve <s_******@yahoo .comwrote:

[...]
So that ChildClass.Clas sName returns "ChildClass " and
ParentClass.Cla ssName returns "ParentClas s"

Since you cannot call the static member (method or property) without
knowing the class name already, how would this benefit you? How is
writing this:

MyStoredData.Ge tData(ChildClas s.ClassName);

better than this:

MyStoredData.Ge tData("ChildCla ss");

or even:

MyStoredData.Ge tData(typeof(Ch ildClass).GetNa me());

?

More generally, while it's not your question specifically, .NET already
has serialization support built in, including being able to serialize to
an XML file. It sounds a little like you might be reinventing the wheel
here.

Pete

[steve]

It helps me because I do not use the class name. I manipulate it to give me an xml token which is based on the class name and I
would have to include this logic in all the methods that need to create the token.

I wrote xml read routine a long time ago. I didn't know about the serialization stuff at that time - or if it existed. But, anyway
the class has a lot of variables which are not always used but I have logic in my read/write routines to allow me to ignore writing
data based on the values in certain fields. I am not sure I could have done this with the serialization that is in the Xml
namespace. It saves me an awful lot of space in the output xml file.

I realized that I did have a ChildClass object around when I was making the call so I made them non-static. I guess I could have
created a static conversion routine and hidden that in the parent class.

I will take a look at the generics stuff.

Thanks to all,
Steve

On Tue, 02 Sep 2008 14:00:05 -0700, "Peter Duniho" <Np*********@nn owslpianmk.comw rote:

>On Tue, 02 Sep 2008 13:33:25 -0700, steve <s_******@yahoo .comwrote:

>[...]
So that ChildClass.Clas sName returns "ChildClass " and
ParentClass.Cl assName returns "ParentClas s"

Since you cannot call the static member (method or property) without
knowing the class name already, how would this benefit you? How is
writing this:

MyStoredData.Ge tData(ChildClas s.ClassName);

better than this:

MyStoredData.Ge tData("ChildCla ss");

or even:

MyStoredData.Ge tData(typeof(Ch ildClass).GetNa me());

?

More generally, while it's not your question specifically, .NET already
has serialization support built in, including being able to serialize to
an XML file. It sounds a little like you might be reinventing the wheel
here.

Pete

[Peter Duniho]

On Tue, 02 Sep 2008 15:43:16 -0700, steve <s_******@yahoo .comwrote:

It helps me because I do not use the class name. I manipulate it to give
me an xml token which is based on the class name and I
would have to include this logic in all the methods that need to create
the token.

Or, you could just create a single static method somewhere that does that
transformation for you, based on an explicitly-given class name. That
logic doesn't have to be in the class, and it might even be more
appropriate for it _not_ to be there (depending on how you've encapsulated
the serialization logic).

If not in the type you're serializing, it could even be an extension
method, either on the Type or String class (depending on what syntax you
prefer).

I wrote xml read routine a long time ago. I didn't know about the
serialization stuff at that time - or if it existed. But, anyway
the class has a lot of variables which are not always used but I have
logic in my read/write routines to allow me to ignore writing
data based on the values in certain fields. I am not sure I could have
done this with the serialization that is in the Xml
namespace. It saves me an awful lot of space in the output xml file.

I get the various serialization types mixed up, so I can't tell you which
one specifically I've used more (such as it is...I don't write a lot of
serialization code). But I can tell you that the one I've used more
allows all sorts of customization and control over the serialization.

Pete

[Marc Gravell]

I am not sure I could have done this with the serialization that is in the Xml

namespace. It saves me an awful lot of space in the output xml file.

Sure you can... for the simple cases you can just specify
[DefaultValue(.. .)]; for more complex cases you just add a bool
ShouldSerialize {prop-name}() method. This method pattern is inherited
from System.Componen tModel, where it works even if private, but IIRC
XmlSerializer likes the method to be public (you can always hide it
from intellisense with EditorBrowsable Attribute). Finally, for the
most extreme cases you can implement IXmlSerializabl e (although this
largely defeats the point).

If you are talking about whitespace etc, then XmlWriterSettin gs is
what you want.

Marc

[steve]

I have been looking at the book "Beginning XML with C# 2008" by Joshi and I can't see anything that would allow me to do what I want
or what you suggest.

A typical example would be writing out a Polygon class where I can give a different color to each edge of the polygon. But, in the
majority of cases all the edges will be the same color or not drawn so that if I write out a node or attribute indicating the edges
do not need to be drawn I know I don't need to write out the attributes for each of the sides.

Also multiple classes get written to the same xml file. Would that be a problem for the serializer?

If you know of a good source to read about this it would be appreciated.

Thanks,
Steve


On Tue, 2 Sep 2008 23:16:11 -0700 (PDT), Marc Gravell <ma**********@g mail.comwrote:

>I am not sure I could have done this with the serialization that is in the Xml
namespace. It saves me an awful lot of space in the output xml file.

Sure you can... for the simple cases you can just specify
[DefaultValue(.. .)]; for more complex cases you just add a bool
ShouldSerializ e{prop-name}() method. This method pattern is inherited
from System.Componen tModel, where it works even if private, but IIRC
XmlSerialize r likes the method to be public (you can always hide it
from intellisense with EditorBrowsable Attribute). Finally, for the
most extreme cases you can implement IXmlSerializabl e (although this
largely defeats the point).

If you are talking about whitespace etc, then XmlWriterSettin gs is
what you want.

Marc

[Marc Gravell]

Also multiple classes get written to the same xml file. Would that be a problem for the serializer?

As long as they are expected, for example via XmlIncludeAttri bute for
sub-types

If you know of a good source to read about this it would be appreciated.

msdn:
http://msdn.microsoft.com/en-us/library/83y7df3e.aspx
http://msdn.microsoft.com/en-us/library/2baksw0z.aspx
http://msdn.microsoft.com/en-us/library/athddy89.aspx

I can't find a good source for ShouldSerialize , but it works: in the
following, Foo is written, Bar isn't:

using System;
using System.Componen tModel;
using System.Xml.Seri alization;

[Serializable, XmlType("test")]
public class Test
{
[XmlAttribute("f oo")]
public string Foo { get; set; }
[EditorBrowsable (EditorBrowsabl eState.Never)]
public bool ShouldSerialize Foo() { return true; }
[XmlAttribute("b ar")]
public string Bar { get; set; }
[EditorBrowsable (EditorBrowsabl eState.Never)]
public bool ShouldSerialize Bar() { return false; }
}

static class Program
{
static void Main()
{
XmlSerializer xser = new XmlSerializer(t ypeof(Test));
Test obj = new Test { Foo = "abc", Bar = "def" };
xser.Serialize( Console.Out, obj);
}
}

[Marc Gravell]

(obviously the ShouldSerialize methods would normally check some other
state to make the decision - for a more interesting example, consider
a rectangle that serializes @width and @height if non-square, and
@size otherwise):

using System;
using System.Componen tModel;
using System.Xml.Seri alization;

[Serializable, XmlType("rectan gle")]
public class Rectangle
{
[XmlAttribute("w idth")]
public int Width { get; set; }
[EditorBrowsable (EditorBrowsabl eState.Never)]
public bool ShouldSerialize Width() { return Width != Height; }

[XmlAttribute("h eight")]
public int Height { get; set; }
[EditorBrowsable (EditorBrowsabl eState.Never)]
public bool ShouldSerialize Height() { return Width != Height; }

[XmlAttribute("s ize")]
public int Size {
get {
if (Width == Height) return Width;
throw new InvalidOperatio nException("Siz e only valid for
squares");
}
set {Width = Height = value;}
}

[EditorBrowsable (EditorBrowsabl eState.Never)]
public bool ShouldSerialize Size() { return Width == Height; }
}

static class Program
{
static void Main()
{
XmlSerializer xser = new XmlSerializer(t ypeof(Rectangle ));
Rectangle rect = new Rectangle { Width = 2, Height = 3 };
xser.Serialize( Console.Out, rect);
rect.Width = 3; // should now be square
xser.Serialize( Console.Out, rect);
}
}