多字符串查找算法：kmp与step

1，计算kmp失败回退函数：

	public static int[] getKmpFail(String pattern) {
		if (!hasLength(pattern)) throw new IllegalArgumentException("null or empty pattern is not allowed to get kmp fail array.");
		
		int i, j, len = pattern.length();
		int[] fail = new int[len];
		fail[0] = -1;
		for (j = 1; j < len; j++) {
			i = fail[j - 1];
			while ((pattern.charAt(j) != pattern.charAt(i + 1)) && (i >= 0)) {
				i = fail[i];
			}
			if (pattern.charAt(j) == pattern.charAt(i + 1)) {
				fail[j] = i + 1;
			} else {
				fail[j] = -1;
			}
		}
		return fail;
	}

２，查找子串：

	public static int kmpIndexOf(String source, String pattern, int from, int[] fail, int to) {
		int i = from, j = 0, lenp = pattern.length();
		while ((i < to) && (j < lenp)) {
			if (source.charAt(i) == pattern.charAt(j)) {
				i++;
				j++;
			} else {
				if (j == 0) {
					i++;
				} else {
					j = fail[j - 1] + 1;
				}
			}
		}
		return (j == lenp) ? (i - lenp) : -1;
	}

3，kmp多字符串查找：

	public static int[] kmpFirstMatch(String source, String[] patterns, int from, int[][] fails) {
		int[] indices = { -1, -1 };
		int to = source.length();
		for (int i = 0; i < patterns.length; i++) {//多次遍历source查找patterns[i]
			if(to - from < patterns[i].length()) continue;
			int index = kmpIndexOf(source, patterns[i], from, fails[i], to);
			if (index != -1 && (index < indices[0] || indices[0] == -1)) {
				indices[0] = to = index;//缩小范围，to之后的不用查找了
				indices[1] = patterns[i].length();//返回目标索引和长度
				to++;//if(from == index) break;上面的continue更优化
			}
		}
		return indices;
	}

4，step多字符串查找：

	public static int[] stepFirstMatch(char[] sources, int from, char[][] chars) {
		int[] indices = { -1, -1 };
		for (int i = from; i < sources.length; i++) {//从头至尾依次查找
			for (int j = 0; j < chars.length; j++) {
				if (sources[i] == chars[j][0]) {
					int k = 1;
					while (k < chars[j].length && i + k < sources.length && sources[i + k] == chars[j][k])
						k++;
					if (k == chars[j].length) {
						indices[0] = i;
						indices[1] = k;
						return indices;//找到就立即返回
					}
				}
			}
		}
		return indices;
	}

step比kmp方式要快一点，大概是基本类型上占些优势。

 

找到stepFirstMatch的一个合适用法：生成随机充值码时排出相似字符串！

//检查相似字符不能同时出现，生成的随机码串调用checkNoSimilarPairs(randomString)即可。
	private static String[] similarPairsString = {"0O", "1I", "2Z", "VY"};
	private static char[][] similarPairs = null;
	static {
		similarPairs = new char[similarPairsString.length*2][];
		for(int i = 0; i < similarPairsString.length; i++) {
			similarPairs[2*i] = similarPairsString[i].toCharArray();
			similarPairs[2*i + 1] = new StringBuilder(similarPairsString[i]).reverse().toString().toCharArray();
		}
	}
	private boolean checkNoSimilarPairs(String source) {
		return stepFirstMatch(source.toCharArray(), 0, similarPairs)[0] == -1;
	}