Recently, I had a need to process the private data contained in the base class using the methods of the child class. My base class could only contain domain-specific types (it only represents data). So first I decided to create a child-class in another project and implement the processing logic in it. But the problem is that once you create an instance of the base class, you can't cast it to the child type:
public class A
{
protected int member1;
public A(int value)
{
member1 = value;
}
}
public class B : A
{
public B (int value) : base(value)
{ }
public void DoSomething()
{
Console.Write(member1 * member1);
}
}
class Program
{
static void Main(string[] args)
{
A obj1 = new A(5);
B obj2 = (B)obj1; // InvalidCastException
obj2.DoSomething();
}
}
And I started thinking towards extension methods. However, you can't just access the protected fields of the class from them. In the end, I tried to combine the two approaches.
Here's my solution:
Make sure that you are allowed to add new methods to your base class and that your class is not sealed.
Add protected static method which returns the protected member you need.
Create an Extension class for your base class.
In extension class create a private nested class.
Inherit your nested class from your base class.
Create static method in nested class and implement the processing logic in (you can call static protected method from base class to get protected member from base class).
Create extension method in extension class and call static method of nested class in it.
The sample code is shown below:
public class A
{
protected int member1 = 0;
public A() {}
public A(int value)
{
member1 = value;
}
protected static int GetProtectedMember(A objA)
{
return objA.member1;
}
}
public static class AExtensions
{
public static void DoSomething(this A objA)
{
B.DoSomething(objA);
}
private class B : A
{
public static void DoSomething(A objA)
{
// objA.member1 // it's not allowed
int protectedFromA = A.GetProtectedMember(objA);
int sqr = protectedFromA * protectedFromA;
Console.WriteLine(sqr);
}
}
}
class Program
{
static void Main(string[] args)
{
A obj1 = new A(5);
obj1.DoSomething(); // 25
}
}
This way you can keep the classes that represent the data in a separate project and have multiple implementations of processing this data in different projects.
Related
I Am getting a warning message in my class, like
Add a Protected constructor or the static keyword to the class declaration
Solution
The error is gone after I tried both the below ways,
static class without constructor
public static class Program {
}
Non static class with protected using constructor
public class Program
{
protected Program() { }
}
Question:
So What is the difference between Static Class vs Protected Constructor which is mentioned in my above solution? And which one is best to use?
A static class doesn't need an instance to access its members. A static class cannot have instance members (e.g. public int MyNumber; is not allowed on a static class because only static members are allowed on a static class). Both instance and static members are allowed on a non-static class though. A class with a protected constructor can only have an instance created by itself or something that inherits from it.
public class Program
{
protected Program()
{
// Do something.
}
public static Program Create()
{
// 100% Allowed.
return new Program();
}
public void DoSomething()
{
}
}
public static class AnotherClass
{
public static Program CreateProgram()
{
// Not allowed since Program's constructor is protected.
return new Program();
}
}
public class SubProgram : Program
{
protected SubProgram()
{
// Calls Program() then SubProgram().
}
public new static Program Create()
{
// return new Program(); // We would need to move the SubProgram class INSIDE the Program class in order for this line to work.
return new SubProgram();
}
}
Program.Create(); // Can be called since Create is public and static function.
Program.DoSomething() // Can't be called because an instance has not been instantiated.
var test = Program.Create();
test.DoSomething(); // Can be called since there is now an instance of Program (i.e. 'test').
AnotherClass.CreateProgram(); // Can't be called since Program's constructor is protected.
SubProgram.Create(); // Can be called since SubProgram inherits from Program.
As for performance, this distinction doesn't really have much to do with performance.
You probably only have static members in the class and the code analyser assumes that your intention is to not be able to create instances of the class so it is asking you to either make the class static
public static class Program {
//...static members
}
or put a protected/private constructor
public class Program {
protected Program { //OR private
}
//...static members
}
to prevent instances of that class from being initialized.
A static class is basically the same as a non-static class, but there is one difference: a static class cannot be instantiated.
Reference Static Classes and Static Class Members (C# Programming Guide)
The protected constructor means that only derived classes can call the constructor
and a private constructor wont allow any other classes to initialize the class with a private constructor
A static constructor is called when the class type is instantiated. The protected constructor is called when an instance of a class is created. The protected part means only classes that inherit the class can call it.
Static Constructor: Called once when the class type is instantiated and is used to initialize static members. Does not create an instance of the class.
Protected Constructor: A constructor that can be called only by the class or a class that inherits it.
The best practices for this is that you should have a static constructor for initializing static members and a protected constructor if you only want classes that inherit to be able to create an instance of your class. You can have both.
public class MyClass
{
static readonly long _someStaticMember;
private bool _param;
static MyClass()
{
//Do Some Logic
_someStaticMember = SomeValueCalculated;
}
protected MyClass(bool param)
{
_param = param;
}
}
public class ChildClass: MyClass
{
public ChildClass(bool param) : base(param);
}
public class NotChildClass
{
public MyClass someObject = new MyClass(true); //Will Fail
}
Can you add a Derived Class to a list of its base class then call a method of the Derived class from the list of base class(possibly by casting it back to the Derived class since you know it was originally a Derived class)
public class MySystem
{
public string name;
MySystem(string name)
{
this.name = name;
}
public void Update()
{
//dostuff
}
}
public class PowerSystem : MySystem
{
public int totalPower;
PowerSystem (string name, int power) : base(name)
{
this.totalPower = power;
}
public void Update()
{
base.Update();
//Do other stuff
}
}
void Main()
{
List<MySystem> SystemList = new List<MySystem>();
SystemList.Add(new System("Shields"));
SystemList.Add(new System("Hull"));
Power p = new Power("Power", 10);
SystemList.Add(p);
foreach(MainSystems ms in SystemList)
{
if(ms.name != "Power")
ms.Update();
else
(PowerSystem)ms.Update(); //This doesn't work
}
}
So what I'm trying to do is run the update method for every element in the list, with the exeption of the one I named power and instead run the Power.Update method.
The closest post I have found to answering this is here unfortunately I don't fully understand it.
I'm hoping that the list is holding a reference to PowerSystem p and that somehow I can convert it and access the PowerSystem menthod.
I hope this is clear.
Thanks
PS if you have a better idea for this I'm all ears.
Use polymorphism - mark Update in base class virtual and override it in derived class.
Base classes may define and implement virtual methods, and derived
classes can override them, which means they provide their own
definition and implementation. At run-time, when client code calls the
method, the CLR looks up the run-time type of the object, and invokes
that override of the virtual method. Thus in your source code you can
call a method on a base class, and cause a derived class's version of
the method to be executed.
public class MySystem
{
public string name;
MySystem(string name)
{
this.name = name;
}
public virtual void Update()
{
//dostuff
}
}
public class PowerSystem : MySystem
{
public int totalPower;
PowerSystem (string name, int power) : base(name)
{
this.totalPower = power;
}
public override void Update()
{
base.Update();
//Do other stuff
}
}
Now, PowerSystem.Update() will get called automatically
foreach(MainSystems ms in SystemList)
{
ms.Update();
}
For MySystem instances it will call MySystem.Update, but for PowerSystem instances the override will be called.
public class Foo
{
public const int type = 1;
}
Why can't i do this? Is there a reason behind it or am I trying to access the constant in a wrong way?
new Foo().type;
I know I can do Foo.type but given my scenario, I cant do that. For example if I have two class which inherit from a base class like this:
public class Base
{
...
}
public class Foo : Base
{
public const int type = 0;
}
public class Bar : Base
{
public const int type = 1;
}
public static void printType(Base b)
{
Console.WriteLine(b.type);
}
I would want to get the type property of the class sent through the printType() function but I cant since I can only access the type from the Class, not the object its self.
A work around would be to do
if(b is Foo){
Console.Write(Foo.type);
}elseif....
but this seems stupid and not viable if you have many sub classes of Base
Solution
I ended up using readonly instead of const like this:
public readonly int type = 0;
Yes, you're trying to access it in the wrong way. A constant isn't associated with an instance of a type - it's associated with the type itself. So you want:
int x = Foo.type;
Basically, const members are implicitly static, and C# doesn't let you access static members as if they were instance members, via a value. (Note that in .NET naming conventions, it should be Type rather than type.)
EDIT: Now that you've explained the actual situation, it appears you're trying to use polymorphism, which won't work for constants. So instead, you should have an abstract property in the base class, implemented in subclasses.
public abstract class Base
{
public abstract int Type { get; }
}
public class Foo : Base
{
public override int Type { get { return 0; } }
}
public class Bar : Base
{
public override int Type { get { return 0; } }
}
Alternatively, just have a normal property in the base class which is populated via the base class constructor:
public class Base
{
private readonly int type;
public int Type { get { return type; } }
protected Base(int type)
{
this.type = type;
}
}
public class Foo : Base
{
public Foo() : base(0) {}
}
public class Bar : Base
{
public Bar() : base(1) {}
}
If you just want something to identify the dynamic (most-derived) type of the object passed in, that's built into .NET, via the Object.GetType() method.
public static void printType(Base b)
{
Console.WriteLine(b.GetType().Name);
}
Of course, this isn't quite the same as having attached data under your control. You can, however, use a Dictionary<Type, T> to associate data of arbitrary type with the various subclasses. It would be reasonable to use the subclass type initializer to install new entries into such a dictionary.
public class Base
{
static internal readonly Dictionary<System.Type, int> TypeMap =
new Dictionary<System.Type, int>();
}
public class Foo : Base
{
static Foo { TypeMap.Add(typeof(Foo), 0); }
}
public class Bar : Base
{
static Bar { TypeMap.Add(typeof(Bar), 1); }
}
public static void printType(Base b)
{
Console.WriteLine(Base.TypeMap[b.GetType()]);
}
This WILL be a bit slower than the field-per-object method, however it doesn't add any extra storage per-object.
There are two files A.cs and B.cs. There is a method fn() which is used in both the classes.
Method fn() is used in both class files. This increases code complexity if I need this method in many class files (say 100 files).
I know that we can call this method by creating an object for the class in which this method is defined. How can I share this function between two or more classes without creating an object every time for accessing this method?
Put the method in a static class:
public static class Utils
{
public static string fn()
{
//code...
}
}
You can then call this in A.cs and B.cs without creating a new instance of a class each time:
A foo = new A();
foo.Property = Utils.fn();
Alternatively, you could create a BaseClass that all classes inherit from:
public class BaseClass
{
public BaseClass() { }
public virtual string fn()
{
return "hello world";
}
}
public class A : BaseClass
{
public A() { }
}
You would then call fn() like so:
A foo = new A();
string x = foo.fn();
I hope the function isn't really called fn(), but actually named to what it does, like CalculateTotal(). Then you can extract this method into a class, say: TotalCalculator.
Now upon application startup, preferably using dependency injection, you create one instance of the class that gets shared between objects that require it. Like so:
class TotalCalculator
{
public int Calculate()
{
return 42;
}
}
class NeedsCalculator1
{
TotalCalculator _calculator;
public NeedsCalculator1(TotalCalculator calculator)
{
_calculator = calculator;
}
public void Foo()
{
_calculator.Calculate();
}
}
class NeedsCalculatorToo
{
TotalCalculator _calculator;
public NeedsCalculatorToo(TotalCalculator calculator)
{
_calculator = calculator;
}
public void Bar()
{
_calculator.Calculate();
}
}
Then you instantiate the calculator once, and pass it into the other classes' constructor:
TotalCalculator calculator = new TotalCalculator();
NeedsCalculator1 dependency1 = new NeedsCalculator1(calculator);
NeedsCalculatorToo dependency2 = new NeedsCalculatorToo(calculator);
You can now further abstract the calculator dependency by creating a base class containing the constructor and a protected TotalCalculator instance field, for example.
Assuming that this method is self contained, you could create a static class and put this method as a static method in it, which is in turn, called by the other classes.
If is is not self contained, you could try and declare it in some super class and let the other 100 classes extend that class.
I have error
Cannot access a non-static member of outer type 'Project.Neuro' via
nested type 'Project.Neuro.Net'
with code like this (simplified):
class Neuro
{
public class Net
{
public void SomeMethod()
{
int x = OtherMethod(); // error is here
}
}
public int OtherMethod() // its outside Neuro.Net class
{
return 123;
}
}
I can move problematic method to Neuro.Net class, but I need this method outside.
Im kind of objective programming newbie.
Thanks in advance.
The problem is that nested classes are not derived classes, so the methods in the outer class are not inherited.
Some options are
Make the method static:
class Neuro
{
public class Net
{
public void SomeMethod()
{
int x = Neuro.OtherMethod();
}
}
public static int OtherMethod()
{
return 123;
}
}
Use inheritance instead of nesting classes:
public class Neuro // Neuro has to be public in order to have a public class inherit from it.
{
public static int OtherMethod()
{
return 123;
}
}
public class Net : Neuro
{
public void SomeMethod()
{
int x = OtherMethod();
}
}
Create an instance of Neuro:
class Neuro
{
public class Net
{
public void SomeMethod()
{
Neuro n = new Neuro();
int x = n.OtherMethod();
}
}
public int OtherMethod()
{
return 123;
}
}
you need to instantiate an object of type Neuro somewhere in your code and call OtherMethod on it, since OtherMethod is not a static method. Whether you create this object inside of SomeMethod, or pass it as an argument to it is up to you. Something like:
// somewhere in the code
var neuroObject = new Neuro();
// inside SomeMethod()
int x = neuroObject.OtherMethod();
alternatively, you can make OtherMethod static, which will allow you to call it from SomeMethod as you currently are.
Even though class is nested within another class, it is still not obvious which instance of outer class talks to which instance of inner class. I could create an instance of inner class and pass it to the another instance of outer class.
Therefore, you need specific instance to call this OtherMethod().
You can pass the instance on creation:
class Neuro
{
public class Net
{
private Neuro _parent;
public Net(Neuro parent)
{
_parent = parent;
}
public void SomeMethod()
{
_parent.OtherMethod();
}
}
public int OtherMethod()
{
return 123;
}
}
I think making an instance of outer class in inner class is not a good option because you may executing business logic on outer class constructor. Making static methods or properties is better option. If you insist making an instance of outer class than you should add another parameter to outer class contructor that not to execute business logic.