19

有谁知道为什么 C# (.NET) 的StartsWith函数比IsPrefix慢很多?

4

4 回答 4

35

我认为它主要是获取线程的当前文化。

如果您更改 Marc 的测试以使用以下形式String.StartsWith

    Stopwatch watch = Stopwatch.StartNew();
    CultureInfo cc = CultureInfo.CurrentCulture;
    for (int i = 0; i < LOOP; i++)
    {
        if (s1.StartsWith(s2, false, cc)) chk++;
    }
    watch.Stop();
    Console.WriteLine(watch.ElapsedMilliseconds + "ms; chk: " + chk);

它更接近了。

如果你使用s1.StartsWith(s2, StringComparison.Ordinal)它比使用快得多CompareInfo.IsPrefix(当然取决于CompareInfo)。在我的盒子上,结果是(不科学的):

  • s1.StartsWith(s2): 6914ms
  • s1.StartsWith(s2, false,culture): 5568ms
  • 比较.IsPrefix(s1, s2): 5200ms
  • s1.StartsWith(s2, StringComparison.Ordinal): 1393ms

显然这是因为它实际上只是在每个点比较 16 位整数,这非常便宜。如果您想要对文化敏感的检查,并且性能对您来说特别重要,那么这就是我要使用的重载。

于 2009-04-04T21:49:42.773 回答
7

好问题; 进行测试,我得到:

9156ms; chk: 50000000
6887ms; chk: 50000000

试验台:

using System;
using System.Diagnostics;
using System.Globalization;    

class Program
{
    static void Main()
    {
        string s1 = "abcdefghijklmnopqrstuvwxyz", s2 = "abcdefg";

        const int LOOP = 50000000;
        int chk = 0;
        Stopwatch watch = Stopwatch.StartNew();
        for (int i = 0; i < LOOP; i++)
        {
            if (s1.StartsWith(s2)) chk++;
        }
        watch.Stop();
        Console.WriteLine(watch.ElapsedMilliseconds + "ms; chk: " + chk);

        chk = 0;
        watch = Stopwatch.StartNew();

        CompareInfo ci = CultureInfo.CurrentCulture.CompareInfo;
        for (int i = 0; i < LOOP; i++)
        {
            if (ci.IsPrefix(s1, s2)) chk++;
        }
        watch.Stop();
        Console.WriteLine(watch.ElapsedMilliseconds + "ms; chk: " + chk);
    }
}
于 2009-04-04T21:43:40.607 回答
6

StartsWith 在内部调用 IsPrefix。它在调用 IsPrefix 之前分配文化信息。

于 2009-04-04T21:39:02.223 回答
1

查看 IsPrefix 的来源。问题是 - 在某些情况下,它会比StartsWith仅仅因为它实际上使用 StartsWith 并且只执行很少的操作而变慢。

    [System.Security.SecuritySafeCritical]  // auto-generated
    public unsafe virtual bool IsPrefix(String source, String prefix, CompareOptions options)
    { 
        if (source == null || prefix == null) {
            throw new ArgumentNullException((source == null ? "source" : "prefix"), 
                Environment.GetResourceString("ArgumentNull_String")); 
        }
        Contract.EndContractBlock(); 
        int prefixLen = prefix.Length;

        if (prefixLen == 0)
        { 
            return (true);
        } 

        if (options == CompareOptions.OrdinalIgnoreCase)
        { 
            return source.StartsWith(prefix, StringComparison.OrdinalIgnoreCase);
        }

        if (options == CompareOptions.Ordinal) 
        {
            return source.StartsWith(prefix, StringComparison.Ordinal); 
        } 

        if ((options & ValidIndexMaskOffFlags) != 0) { 
            throw new ArgumentException(Environment.GetResourceString("Argument_InvalidFlag"), "options");
        }


        // to let the sorting DLL do the call optimization in case of Ascii strings, we check if the strings are in Ascii and then send the flag RESERVED_FIND_ASCII_STRING  to
        // the sorting DLL API SortFindString so sorting DLL don't have to check if the string is Ascii with every call to SortFindString. 

        return (InternalFindNLSStringEx(
                    m_dataHandle, m_handleOrigin, m_sortName, 
                    GetNativeCompareFlags(options) | Win32Native.FIND_STARTSWITH | ((source.IsAscii() && prefix.IsAscii()) ? RESERVED_FIND_ASCII_STRING : 0),
                    source, source.Length, 0, prefix, prefix.Length) > -1);
    }
于 2013-06-22T08:32:04.900 回答