file3 (單詞1，單詞a，單詞3，單詞d....)

...  

下面是我對於倒排索引的一個簡單的實現。該程式實現了某目錄下的檔案單詞統計，並建立了倒排索引。

標頭檔案：inverted_index.h

#define MAXLINE 1024

class InvertedIndex
{
public:
	InvertedIndex(){}
	~InvertedIndex(){result_index_.clear();}

	int StatWords(const char* file_name);
	map<string,map<string,int> > result_index(){return result_index_;}
	

private:
	//存放倒排索引結果，key是單詞，value也是map，該map的key是檔名，value是該單詞在該檔案中出現的次數
	map<string,map<string,int> > result_index_;
	
	int ParseWordsByLine(char* str_line, const char* file_name);
	void InsertWordToMap(char* word, const char* file_name);
};
 

實現檔案：inverted_index.cpp

int InvertedIndex::StatWords(const char* file_name)
{
	FILE *fp;   
    if((fp = fopen(file_name,"r")) == NULL)
    { 
        printf("cannot open %s", file_name);
        return -1;
    }
    
    char str_line[MAXLINE];
	while(fgets(str_line, MAXLINE, fp) != NULL)
	{
		int len = strlen(str_line);
		str_line[len-1] = '\0';  /*去掉換行符*/
		ParseWordsByLine(str_line,file_name);
	} 
    fclose(fp);
    return 0; 
}


int InvertedIndex::ParseWordsByLine(char* str_line, const char* file_name)
{
	if(strlen(str_line) == 0)
	{
		return -1;
	}

	const char* needle=" \n\t";
	/*strtok在s中查詢包含在delim中的字元並用NULL(‘\0’)來替換，直到找遍整個字串。 
	返回值：從s開頭開始的一個個被分割的串。當沒有被分割的串時則返回NULL。*/
	char* word = strtok(str_line, needle);
	while(word != NULL)
	{
		InsertWordToMap(word,file_name);
	    word = strtok(NULL, needle);
	}
	return 0;
}

void InvertedIndex::InsertWordToMap(char* word, const char* file_name)
{
	if(strlen(word) == 0)
	{
		return;
	}
	string word_str = word;
	string file_name_str = file_name;

	map<string,map<string,int> >::iterator it;
	it = result_index_.find(word_str);

	if(it == result_index_.end())//not found
	{
		map<string,int> file_word_count;
		file_word_count.insert(pair<string,int>(file_name_str,1));
		result_index_[word_str] = file_word_count;
	}
	else
	{
		map<string,int> file_word_count = result_index_[word_str];
		file_word_count[file_name_str] ++ ;
		result_index_[word_str] = file_word_count;
	}
}