如何在.NET中将自定义类型保存为二进制数据？

Question

I'm creating something that has save files. For the most part, it's just saving objects with 4 values.

• x
• y
• id
• layer

I've been saving this as XML for a while, the problem is, the files are starting to get HUGE because I'm saving hundreds to thousands of these objects and XML is kind of bulky for something like this. I tried switching to Json but it was too big as well (I will admit though, better).

My Question

I know a lot of programs save directly using bytes to conserve space. I want to do this. Say I have 300 objects with the properties X, Y, Id, and Layer, how could I save this to file x as bytes and load it later?

I've tried reading bytes when I used to make my own servers. I usually ended up getting frustrated and giving up. I'm hoping this isn't too similar (my gut says otherwise).

EDIT

Oh sorry guys, I forgot to mention, I'm using VB.NET so all .NET answers are acceptable! :)

Also, all of these values are integers.

Answer 1

I would use protobuf-net (I would; I'm biased...). It is an open-source .NET serializer, that uses attributes (well, that is optional in v2) to guide it. For example, using C# syntax (purely for my convenience - the engine doesn't care what language you use):

[ProtoContract]
public class Whatever {
    [ProtoMember(1)]
    public int X {get;set;}
    [ProtoMember(2)]
    public int Y {get;set;}
    [ProtoMember(3)]
    public long Id {get;set;}
    [ProtoMember(4)]
    public string Layer {get;set;}
}

You can then serialize, for example:

List<Whatever> data = ...
Serializer.Serialize(file, data);

and deserialize:

var data = Serializer.Deserialize<List<Whatever>>(file);

job done; fast binary output. You probably can get a little tighter by hand-coding all reading/writing manually without any markers etc, but the above will be a lot less maintenane. It'll also be easy to load into any other platform that has a "protocol buffers" implementation available (which is: most of them).

Answer 2

Alternative: why not xml and zip (or some other compression)

Advantage:
* its still "human" readable
* text files have normaly a good compression factor

Also what programming languae you are using? in C/C++ or you create a struct with the data reinterpret cast the struct to byte * and write sizeof bytes to the file... simple as that (assuming the struct is fixed size (id and layer are also fixed sized variables like int)

Answer 3

为什么不使用像MySql或SQL这样的数据库？