Question

通过数组（或列表，linkedList，Dictionary ect）的迭代速度是否取决于数据类型？

实施例：一个10个bool v / s的数组，一个10个整数的数组？

Answer 1

是的，数据类型很重要。它与迭代无关;它与数据类型有关。

值类型

int的长度为4个字节。 decimal的长度为16个字节。因此，decimal比int大4倍。每次从数组中检索值时，都会复制该值。如果decimal，则复制16个字节。（如果是引用类型，则复制引用，通常为4或8个字节）。复制更多字节只会减慢迭代速度。

<强>拳击

如果您通过集合进行迭代，则可能还有更改类型。例如：

foreach(object o in new int[] { 1,2,3 })
     ....

这会将每int个框添加到object。这需要时间。这与迭代无关，它与你拳击的事实有关。

<强>铸造

最后一个例子：您还必须投射数组：

foreach(Person p in new object[] { ... })
     ....

施法也需要额外的时间。

编辑

用于备份我的声明的一些时间测量：

Times in milliseconds. Arrays are of size 10,000. Iterations also 10,000.

Answer 2

如果需要，请运行以下代码，但这里是一个快速比较。它所做的就是迭代数组/列表，并将temp变量设置为该索引中的值。

3rd run, high priority Threw some more types in

请注意，现在跑步时Int表现不知所措......不知道为什么......但它也会在重复运行中发生......

    namespace Iterating_types
    {
    using System;
    using System.Collections.Generic;
    using System.Diagnostics;
    using System.Linq;
    using System.Text;
class Program
    {
        static void Main(string[] args)
        {
            Thread.CurrentThread.Priority = ThreadPriority.Highest;
            Process.GetCurrentProcess().PriorityClass = ProcessPriorityClass.RealTime;

            Stopwatch watch = new Stopwatch();
            int UPPER = 1000000;
            int[] int_arr = Enumerable.Range(1, UPPER).ToArray();
            List<int> int_list = Enumerable.Range(1, UPPER).ToList();

            Int32[] int32_arr = Enumerable.Range(1, UPPER).ToArray();

            Int64[] int64_arr = new Int64[UPPER]; 

            IntObject[] intobject_arr = new IntObject[UPPER];
            List<IntObject> intobject_list = new List<IntObject>();

            string[] string_arr = new string[UPPER];
            List<string> string_list = new List<string>();

            bool[] bool_arr = new bool[UPPER];
            Boolean[] boolean_arr = new Boolean[UPPER];
            List<bool> bool_list = new List<bool>();
            List<Boolean> boolean_list = new List<Boolean>();
            // Initializing some of the above
            for (int i = 0; i < UPPER; i++)
            {
                int64_arr[i] = (Int64) i;
                string_arr[i] = i.ToString();
                string_list.Add(i.ToString());
                intobject_arr[i] = new IntObject(i);
                intobject_list.Add(new IntObject(i));
                bool_arr[i] = (i%2 ==0);
                boolean_arr[i] = (i%2 ==0);
                bool_arr[i] = (i%2 ==0);
                bool_list.Add(i%2 ==0);

                boolean_list.Add(i%2 == 0);
            }

            Console.WriteLine("Iterations: {0}{1}", UPPER, Environment.NewLine);
            Console.WriteLine("Thread priority: {0}", Thread.CurrentThread.Priority);
            Console.WriteLine("Process priority: {0}", Process.GetCurrentProcess().PriorityClass);

            Console.WriteLine("\n\rArrays:\t----------------------------------------------");

            bool b;
            b = bool_arr[1];
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                b = bool_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: bool\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                b = boolean_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: Boolean\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            int temp_int;
            temp_int = int_arr[1];
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                temp_int = int_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: Int\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            Int32 temp_int32 ;
            temp_int32 = int32_arr[1];
            watch.Reset();
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                temp_int32 = int32_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: Int32\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            Int64 temp_int64 ;
            temp_int64 = int64_arr[1];
            watch.Reset();
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                temp_int64 = int64_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: Int64\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            string s ;
            s = string_arr[1];
            watch.Reset();
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                s = string_arr[i];
            }
            watch.Stop();
            Console.WriteLine("Type: string\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            temp_int = intobject_arr[1].IntValue;
            watch.Reset();
            watch.Start();
            for (int i = 0; i < UPPER; i++)
            {
                temp_int = intobject_arr[i].IntValue;
            }
            watch.Stop();
            Console.WriteLine("Type: IntObject\tStructure: Array\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            Console.WriteLine("\n\rLists:\t----------------------------------------------");

            watch.Reset();
            watch.Start();
            foreach (var val in bool_list)
            {
                b = val;
            }
            watch.Stop();
            Console.WriteLine("Type: bool\tStructure: List\t\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            watch.Reset();
            watch.Start();
            foreach (var val in boolean_list)
            {
                b = val;
            }
            watch.Stop();
            Console.WriteLine("Type: Boolean\tStructure: List\t\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            temp_int = int_list.First();
            watch.Reset();
            watch.Start();
            foreach (var val in int_list)
            {
                temp_int = val;
            }
            watch.Stop();
            Console.WriteLine("Type: Int\tStructure: List\t\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            s = string_list.First();
            watch.Reset();
            watch.Start();
            foreach (var val in string_list)
            {
                s = val;
            }
            watch.Stop();
            Console.WriteLine("Type: string\tStructure: List\t\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            temp_int = intobject_list.First().IntValue;
            watch.Reset();
            watch.Start();
            foreach (var val in intobject_list)
            {
                temp_int = val.IntValue;
            }
            watch.Stop();
            Console.WriteLine("Type: IntObject\tStructure: List\t\tticks: {0}\tMiliSeconds:{1}", watch.ElapsedTicks, watch.ElapsedMilliseconds);

            Console.WriteLine();
            Console.WriteLine("Hit any key to exit.");
            Console.ReadKey();


        }
    }

    class IntObject
    {
        public int IntValue { get; set; }

        public IntObject ()
        {
            IntValue = 0;
        }

        public IntObject(int i)
        {
            IntValue = i;
        }
    }
}

Answer 3

对于参考类型，简单的答案是是，对于值类型，是。

这是因为.NET泛型的实现是以boxing/unboxing is avoided when using Value Types的方式完成的，尽管不在ArrayLists中。例如，List<int>将数组整数直接存储为堆上的整数而不是对象。在参考类型的情况下，例如List<string>，List<person>但是，从对象到数据类型的转换/转换会有一点时间损失。

请参阅comparison between HashSet and List using strings and objects。

当您进行大量迭代时，决定在List，LinkedList，Dictionary，HashSet等之间使用哪一个 >主要是了解它们的存储方式及其运行时间的复杂性。下面列出了一些.NET泛型的实现和渐近索引/迭代时间复杂度：

.NET泛型的内部实现/渐近时间复杂度

+------------------+---------------------------------------+-------------------------+-------------+ | | | Item[i] | | | Name | Internal Implementation |------------+------------| Iteration | | | | Avg. Case | Worst Case | | +------------------+---------------------------------------+------------+------------+-------------+ | List | Array | O(1) | O(1) | O(n) | | LinkedList | Doubly Linked List | O(n) | O(n) | O(n) | | Dictionary | Hashtable with links to another array | O(1) | O(n) | O(n) | | HashSet | Hashtable with links to another array | O(1) | O(n) | O(n) | | SortedDictionary | Red-black tree | O(log n) | O(log n) | O(n) | | SortedList | Array | O(1) | O(n) | O(n) | | SortedSet | Red-black tree | O(n) | O(n) | O(n) | +------------------+---------------------------------------+------------+------------+-------------+

总结，可以根据时间复杂度确定迭代这些数据类型的最可能速度。就快速查找项目而言，List，SortedList，Dictionary和HashSet将始终胜过其他项目，List和{{SortedList如果您要处理大量项目然后将 Dictionary和HashSet置于大型列表的优势（其中性能），则不建议使用1}}最重要的。）

<强>参考文献：

Runtime Complexity of .NET Generic Collection

Comparative Analysis of List, HashSet and SortedSet

Time complexity overview: Dictionary classes

Generic Collections in C# - Time complexity overview

Algorithms: Big-Oh Notation

Sorted Dictionary(Tkey, Tvalue) - MSDN

List(T) - MSDN

<强>词汇表：

Red-black tree

Boxing and Unboxing

迭代速度和数据类型

3 个答案:

.NET泛型的内部实现/渐近时间复杂度