Question

一旦我将所有C驱动器编入索引进行快速搜索，它就会帮助很多线程搜索文件！唯一的问题是我找到了重复的文件。如果有人能解释我为什么在以下算法中有重复文件，我将不胜感激。

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading;
using System.Threading.Tasks;

class Program
{
    static volatile List<System.IO.FileSystemInfo> files = new List<System.IO.FileSystemInfo> ( );
    static readonly object lockFiles = new object ( );  // every time we read or write to files we use this lock

    static long numberOfThreadsRuning; // used to determine when we are done. if 0 threads runing it means we are done!

    static readonly int max_Number_OF_Threads = 8; 

    static void Main (  )
    {
        numberOfThreadsRuning = 0;

        // add first task this will increase numberOfThreadsRuning
        AddTask ( () => findFiles ( @"C:\Users\Antonio" ) );

        // while number of threads running is greater than 0
        while ( Interlocked.Read ( ref numberOfThreadsRuning ) > 0 )   
            Thread.Sleep ( 100 );            
    }

    // start finding files in a new thread
    static void AddTask (Action task )
    {
        // increment numberOfThreadsRuning with a lock
        Interlocked.Increment ( ref numberOfThreadsRuning );

        Task.Factory.StartNew ( task ).ContinueWith ( x =>
        {
            // once we are done executing the task decrement number of threads runing
            Interlocked.Decrement ( ref numberOfThreadsRuning );
        });
    }

    // recursively start finding files
    static void findFiles ( string path )
    {
        System.IO.FileSystemInfo[ ] subDirAndFiles;

        try {
            subDirAndFiles = new System.IO.DirectoryInfo (path).GetFileSystemInfos ( );
        }
        catch { // sometimes we are not able to access some directories so just ignore the error
            return;
        }

        foreach ( var file in subDirAndFiles )
        {
            lock(lockFiles)
                files.Add ( file );

            if ( file is System.IO.DirectoryInfo ) // if it is a directory
            {
                var numTaskRun = Interlocked.Read (ref  numberOfThreadsRuning );

                if ( numTaskRun < max_Number_OF_Threads )
                    AddTask ( ( ) => findFiles ( file.FullName ) ); // if there are 8 or less threads running create a new task
                else
                    findFiles ( file.FullName ); // otherwise continue finding files in current thread
            }                
        }
    }
}

请尝试通过将静态变量max_Number_OF_Threads更改为1来比较使用8个线程VS 1时的速度。证明我的算法错误的另一种方法是，每次我有超过1个线程时，它总会找到不同数量的文件(files.Length = different number)。

Answer 1

这是问题所在：

foreach ( var file in subDirAndFiles )
{

    ...
    AddTask ( ( ) => findFiles ( file.FullName ) );
}

（除了大量的空间，这对我来说真的很难读......还有非传统的方法名称。）

您正在捕获lambda表达式中的循环变量。这意味着（在C＃3和4中）当lambda表达式执行时，将对进行求值 - 这意味着您可以从以后的迭代中获取值。

修复方法是将值复制到循环内的“new”变量中：

foreach (var file in subDirAndFiles) { var copy = file; AddTask(() => FindFiles(copy.FullName)); }

这在C＃5编译器中已经改变，因此您不需要“复制”变量 - 相反，就好像foreach在每次迭代时声明一个新变量，这是一种更自然的方法

为什么递归算法会添加重复文件？

1 个答案: