这件事情似乎很无聊，但是没人做，我来做下苦力吧。
以下是一些简单的测试。单位以ms计算。注意里面用到循环的数量有些事1W，有些是100W。
这些记录可以作为编程过程中的一些参考。
测试平台：
奔腾1.6G 双核CPU
1G内存
vs2008 调试环境测试。
一、ArrayList (100W,1W)
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.ArrayList al = new System.Collections.ArrayList();
for (int i = 0; i < 1000000; i++)
{
al.Add(i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 10000; i++)
{
al.Contains(i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
98 775
160 891
107 773
193 769
二、Hashtable (100W,1W)
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.Hashtable ht = new System.Collections.Hashtable();
for (int i = 0; i < 1000000; i++)
{
ht.Add(i, i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 10000; i++)
{
ht.ContainsKey(i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
}
375 0
673 0
540 0
495 0
把timer1提高到100万(Hashtable (100W,100W))
389 139
616 277
516 140
610 277
三、HashSet (100W,100W)
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.Generic.HashSet<int> ht = new System.Collections.Generic.HashSet<int>();
for (int i = 0; i < 1000000; i++)
{
ht.Add(i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 1000000; i++)
{
ht.Contains(i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
89 32
79 32
79 32
117 31
四、List (100W,1W)
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.Generic.List<int> ht = new System.Collections.Generic.List<int>();
for (int i = 0; i < 1000000; i++)
{
ht.Add(i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 10000; i++)
{
ht.Contains(i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
16 379
19 392
18 403
18 392
把List<int>换成List<object>
96 945
157 1033
106 909
193 910
换成string,i.ToString()
496 1238
531 1190
572 1246
536 1258
五、Dictionary (100W,100W)
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.Generic.Dictionary<int, int> ht = new System.Collections.Generic.Dictionary<int, int>();
for (int i = 0; i < 1000000; i++)
{
ht.Add(i, i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 1000000; i++)
{
ht.ContainsKey(i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
113 35
125 34
124 34
126 34
六、Dictionary Linq查询 (100W,100W)
用个Linq试试
Stopwatch timer = new Stopwatch();
timer.Start();
System.Collections.Generic.Dictionary<int, int> ht = new System.Collections.Generic.Dictionary<int, int>();
for (int i = 0; i < 1000000; i++)
{
ht.Add(i, i);
}
timer.Stop();
Console.Write(timer.ElapsedMilliseconds.ToString() + "\t");
Stopwatch timer1 = new Stopwatch();
timer1.Start();
for (int i = 0; i < 1000000; i++)
{
ht.Where(c => c.Key == i);
}
timer1.Stop();
Console.WriteLine(timer1.ElapsedMilliseconds.ToString());
112 177
107 78
125 70
107 82
结论：
1、如果是使用缓存的话，那么3.5带来的单泛型集合的HashSet可以替代List了。虽然载入速度慢一点，但是查询速度要比List泛型快很多。要注意到，上述测试List的查询时万级的，而HashSet是百万级
2、Dictionary泛型可以替换掉Hashtable了，虽然如果在字符或者object类型下可能会没这么明显。但是在数字类型的匹配上，Dictionary比Hashtable大概快了2倍，而这个开销估计是Hashtable的装箱造成的。
3、Linq还是要慢一些。
by yurow. http://www.cnblogs.com/birdshover/