Return an empty IEnumerator

百般思念 提交于 2019-11-30 06:01:43

This is simple in C# 2:

public IEnumerator GetEnumerator()
{
    yield break;
}

You need the yield break statement to force the compiler to treat it as an iterator block.

This will be less efficient than a "custom" empty iterator, but it's simpler code...

user287107

There is an extra function in the framework:

public static class Enumerable
{
    public static IEnumerable<TResult> Empty<TResult>();
}

Using this you can write:

var emptyEnumerable = Enumerable.Empty<int>();
var emptyEnumerator = Enumerable.Empty<int>().GetEnumerator();

You could implement a dummy class that implements IEnumerator, and return an instance of it:

class DummyEnumerator : IEnumerator
{
    public object Current
    {
        get
        {
            throw new InvalidOperationException();
        }
    }

    public bool MoveNext()
    {
        return false;
    }

    public void Reset()
    {
    }
}

I was curious and went a bit further. I made a test that checks how efficient the methods are comparing yield break, Enumerable.Emtpy and custom class.

You can check it out on dotnetfiddle https://dotnetfiddle.net/vTkmcQ or use the code below.

The result of one of the many dotnetfiddle runs using 190 000 iterations was:

Yield break: 00:00:00.0210611

Enumerable.Empty(): 00:00:00.0192563

EmptyEnumerator instance: 00:00:00.0012966

using System;
using System.Diagnostics;
using System.Collections;
using System.Linq;

public class Program
{
    private const int Iterations = 190000;
    public static void Main()
    {
        var sw = new Stopwatch();

        sw.Start();
        for (int i = 0; i < Iterations; i++)
        {
            IEnumerator enumerator = YieldBreak();
            while(enumerator.MoveNext())
            {
                throw new InvalidOperationException("Should not occur");
            }           
        }
        sw.Stop();

        Console.WriteLine("Yield break: {0}", sw.Elapsed);

        GC.Collect();

        sw.Restart();
        for (int i = 0; i < Iterations; i++)
        {
            IEnumerator enumerator = Enumerable.Empty<object>().GetEnumerator();
            while(enumerator.MoveNext())
            {
                throw new InvalidOperationException("Should not occur");
            }           
        }
        sw.Stop();

        Console.WriteLine("Enumerable.Empty<T>(): {0}", sw.Elapsed);

        GC.Collect();

        sw.Restart();
        var instance = new EmptyEnumerator();
        for (int i = 0; i < Iterations; i++)
        {
            while(instance.MoveNext())
            {
                throw new InvalidOperationException("Should not occur");
            }           
        }
        sw.Stop();

        Console.WriteLine("EmptyEnumerator instance: {0}", sw.Elapsed);
    }

    public static IEnumerator YieldBreak()
    {
        yield break;
    }

    private class EmptyEnumerator : IEnumerator
    {
        //public static readonly EmptyEnumerator Instance = new EmptyEnumerator();

        public bool MoveNext()
        {
            return false;
        }

        public void Reset()
        {
        }

        public object Current { get { return null; } }
    }
}

The way I use is to use the enumerator of an empty array:

public IEnumerator GetEnumerator() {
    return new object[0].GetEnumerator();
}

It can also be used for generic IEnumerator or IEnumerable (use an array of the appropriate type)

You can implement IEnumerator interface and IEnumerable, and return false from MoveNext function of IEnumerable interfase

private class EmptyEnumerator : IEnumerator
{


    public EmptyEnumerator()
    {
    }

    #region IEnumerator Members

    public void Reset() { }

    public object Current
    {
        get
        {
            throw new InvalidOperationException();
        }
    }
    public bool MoveNext()
    { return false; }
}


public class EmptyEnumerable : IEnumerable
{

    public IEnumerator GetEnumerator()
    {
        return new EmptyEnumerator();
    }
}

I wrote it like this:

public IEnumerator<T> GetEnumerator()
{
    return this.source?.GetEnumerator() ??
            Enumerable.Empty<T>().GetEnumerator();
}

You can make a NullEnumerator which implements the IEnumerator interface. You can just pass an instance off the NullEnumerator.

here is an example of an EmptyEnumerator

Found this question looking for the simplest way to get an empty enumerator. After seeing the answer comparing performance I decided to use the empty enumerator class solution, but mine is more compact than the other examples, and is a generic type, and also provides a default instance so you don't have to create new instances all the time, which should even further improve performance.

class EmptyEnumerator<T> : IEnumerator<T>
{
   public readonly static EmptyEnumerator<T> value = new EmptyEnumerator<T>();
   public T Current => throw new InvalidOperationException();
   object IEnumerator.Current => throw new InvalidOperationException();
   public void Dispose() { }
   public bool MoveNext() => false;
   public void Reset() { }
}
标签
易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!