这是我对BMH算法的实现(它就像一个魅力):
public static Int64 IndexOf(this Byte[] value,Byte[] pattern) { if (value == null) throw new ArgumentNullException("value"); if (pattern == null) throw new ArgumentNullException("pattern"); Int64 valueLength = value.LongLength; Int64 patternLength = pattern.LongLength; if ((valueLength == 0) || (patternLength == 0) || (patternLength > valueLength)) return -1; Int64[] badCharacters = new Int64[256]; for (Int64 i = 0; i < 256; ++i) badCharacters[i] = patternLength; Int64 lastPatternByte = patternLength - 1; for (Int64 i = 0; i < lastPatternByte; ++i) badCharacters[pattern[i]] = lastPatternByte - i; // Beginning Int64 index = 0; while (index <= (valueLength - patternLength)) { for (Int64 i = lastPatternByte; value[(index + i)] == pattern[i]; --i) { if (i == 0) return index; } index += badCharacters[value[(index + lastPatternByte)]]; } return -1; }
我试图修改它以返回所有匹配而不是仅返回第一个索引,但我到处都得到IndexOutOfRangeException D:
显然我错过了一些重要的东西,或者我没有正确理解它是如何工作的.我究竟做错了什么?
public static List<Int64> IndexesOf(this Byte[] value,Byte[] pattern) { if (value == null) throw new ArgumentNullException("value"); if (pattern == null) throw new ArgumentNullException("pattern"); Int64 valueLength = value.LongLength; Int64 patternLength = pattern.LongLength; if ((valueLength == 0) || (patternLength == 0) || (patternLength > valueLength)) return (new List<Int64>()); Int64[] badCharacters = new Int64[256]; for (Int64 i = 0; i < 256; ++i) badCharacters[i] = patternLength; Int64 lastPatternByte = patternLength - 1; for (Int64 i = 0; i < lastPatternByte; ++i) badCharacters[pattern[i]] = lastPatternByte - i; // Beginning Int64 index = 0; List<Int64> indexes = new List<Int64>(); while (index <= (valueLength - patternLength)) { for (Int64 i = lastPatternByte; value[(index + i)] == pattern[i]; --i) { if (i == 0) indexes.Add(index); } index += badCharacters[value[(index + lastPatternByte)]]; } return indexes; }
解决方法
更改
if (i == 0) indexes.Add(index);
至
if (i == 0) { indexes.Add(index); break; }