merghash.h

此代码运行于visual c++ 6.0的环境下

// MergHash.h: interface for the CMergHash class.
//
//////////////////////////////////////////////////////////////////////

#if !defined(AFX_MERGHASH_H__BC287431_DFA6_41EB_A1E5_458B62924038__INCLUDED_)
#define AFX_MERGHASH_H__BC287431_DFA6_41EB_A1E5_458B62924038__INCLUDED_

#if _MSC_VER > 1000
#pragma once
#endif // _MSC_VER > 1000


#include "MergeHashElem.h"
#include <fstream>
#include <iostream>
using std::ios;
using std::cout;

using std::fstream;
using std::endl;

#define Max 100000

class CMergHash  
{
public:
	
	CDeque<CMergeHashElem> * Index;
	int ELFHash( const char * key ){
		unsigned long h = 0;
		while( * key ){
			h = ( h << 4 ) + *key++;
			unsigned long g = h & 0xF0000000L;
			if(g) h^=g>>24;
			h &= ~g;
		}
		return h%Max;
	}
	

	void insert( const CMergeHashElem & elem ){
		int num = ELFHash( elem.key.c_str() );
			for( int i = 0 ;i < Index[num].size(); i++ )
			{
				if( Index[num].at(i).key.compare( elem.key ) == 0 )
				{
					Index[num].at(i).pos.push_back( elem.pos.at(0));
					Index[num].at(i).name.push_back( elem.name.at(0));
					return;
				}
			}
			Index[num].push_back( elem );
	}
	

	void visit( CDeque<CMergeHashElem> & Deque, fstream & out ){
		
		for( int i = 0; i < Deque.size(); i++ )//对每一个关键码进行操作
		{
			int flag = 0;//标记是不是第一次
			int sum = 0;//记录索引的个数
		
			for( int k = 0; k < Deque.at(i).pos.size(); k++ )
			{//统计在所有倒排文件中关键码对应的索引的个数
				//if( Deque.at(k).pos.size() != Deque.at(k).name.size() )
				//cout<<"error";
				char temp[100];//name[100];
				int num = 0;
				//_itoa( k+1, name,10 );
				//strcat( name,".txt");
				fstream add;
				add.open(Deque.at(i).name.at(k).c_str(),ios::in);
				//cout<<Deque.at(i).name.at(k).c_str()<<endl;
				add.seekg( Deque.at(i).pos.at(k) );//找到关键码在倒排文件中的位置
				add >>temp;
				add >>num;
				sum+=num;//个数累加
				add.close();
			}
			for( int j = 0; j < Deque.at(i).pos.size(); j++ )
			{
				char temp[100];
				long num,pos;
			//	_itoa( j+1, name,10 );
			//	strcat( name,".txt");
				fstream read;
				read.open(Deque.at(i).name.at(j).c_str(),ios::in);
				
				read.seekg( Deque.at(i).pos.at(j) );
				read >>temp;//读关键码
				if( flag == 0 )//只显示一次
				{
					out<<endl;
					out <<temp<<' ';
					out<<sum;//显示个数
				}
				read >> num;
				for( int k = 0 ; k < num; k++ )
				{
					read >>temp;//读文件名
					out <<endl;
					out <<temp<<' ';//输出文件名
					read >> pos;//读偏移量
					out <<pos;//输出偏移量
				}
				flag = 1;//标记关键码已经输出
				read.close();
			}
		}
	}
	
	void VisitAll(fstream & out){
		for( int i = 0; i < Max; i++ )
			visit( Index[i],out);
	}


	CMergHash(){ Index = new CDeque<CMergeHashElem>[Max]; }
	virtual ~CMergHash(){ delete [] Index; }

};
#endif // !defined(AFX_MERGHASH_H__BC287431_DFA6_41EB_A1E5_458B62924038__INCLUDED_)