现在过虑敏感字算法很多，五花八门的，但很多的不是把敏感字替换成‘*’就是直接消灭了了事，但我现在想要的是替换成特定的字符串。
a.txt 中设定的关键字替换：
尼玛的｜{ 编者按：[你太坏了]}
我就太阳了｜{编者按：[不要这样子]}
F.U.C.K｜{[别骂人，这样真的不好。]}
去si吧｜{[你确定要这样吗？]}
你了不起｜{[好吧，败给你了。一会带你看星星！]}
现在我想把‘｜’前面定义为[敏感字]，然后替换成花括号‘{}’中的字符串。
有没有类似于二维数组这样的存储效果：
string[][] array = new string[][];
array[0][0] = "尼玛的";
array[0][1] = "编者按：[你太坏了]";
array[1][0] = "我就太阳了";
array[1][1] ="编者按：[不要这样子]";
一碰到array[i][0]的敏感字，就替换为array[i][1]中的字符串。
http://www.cnblogs.com/yeerh/archive/2011/10/20/2219035.html 优化的算法
/// <summary>
/// 优化的算法
/// </summary>
public class BadWordsFilter2
{
private HashSet<string> hash = new HashSet<string>();
private ushort[] fastCheck = new ushort[char.MaxValue + 1];
private ushort[] startLength = new ushort[char.MaxValue + 1];
private ushort[] endLength = new ushort[char.MaxValue + 1];
private int maxWordLength = 0;
private int minWordLength = int.MaxValue;
public void AddKey(string word)
{
if (word.Length > 16)
{
throw new Exception("参数最大16个字符");
}
maxWordLength = Math.Max(maxWordLength, word.Length);
minWordLength = Math.Min(minWordLength, word.Length);
//字符出现的位置(1-16),
for (int i = 0; i < word.Length; i++)
{
fastCheck[word[i]] |= (byte)(1 << i);
}
ushort mask = (ushort)(1 << word.Length - 1);
//以x开始的字符的长度
startLength[word[0]] |= mask;
//以x结束的字符的长度
endLength[word[word.Length - 1]] |= mask;
hash.Add(word);
}
public bool HasBadWord(string text)
{
for (int index = 0; index < text.Length; index++)
{
int count = 0;
int maxIndex = Math.Min(maxWordLength + index, text.Length);
char begin = text[index];
for (int j = index; j < maxIndex; j++)
{
char current = text[j];
ushort mask = (ushort)(1 << count);
if ((fastCheck[current] & mask) == 0)
{
index += count;
break;
}
++count;
if ((startLength[begin] & mask) > 0 && (endLength[current] & mask) > 0)
{
string sub = text.Substring(index, count);
if (hash.Contains(sub))
{
//index += (count - 1);
return true;
}
}
}
}
return false;
}
public string FindOne(string text)
{
for (int index = 0; index < text.Length; index++)
{
int count = 0;
int maxIndex = Math.Min(maxWordLength + index, text.Length);
char begin = text[index];
for (int j = index; j < maxIndex; j++)
{
char current = text[j];
ushort mask = (ushort)(1 << count);
if ((fastCheck[current] & mask) == 0)
{
index += count;
break;
}
++count;
if ((startLength[begin] & mask) > 0 && (endLength[current] & mask) > 0)
{
string sub = text.Substring(index, count);
if (hash.Contains(sub))
{
index += (count - 1);
return sub;
}
}
}
}
return string.Empty;
}
public IEnumerable<string> FindAll(string text)
{
for (int index = 0; index < text.Length; index++)
{
int count = 0;
int maxIndex = Math.Min(maxWordLength + index, text.Length);
char begin = text[index];
for (int j = index; j < maxIndex; j++)
{
char current = text[j];
ushort mask = (ushort)(1 << count);
if ((fastCheck[current] & mask) == 0)
{
index += count;
break;
}
++count;
if ((startLength[begin] & mask) > 0 && (endLength[current] & mask) > 0)
{
string sub = text.Substring(index, count);
if (hash.Contains(sub))
{
index += (count - 1);
yield return sub;
break;
}
}
}
}
}
}
