xml2dbtool.java

用来为垂直搜索引擎抓取数据的采集系统

/*
 * *****************************************************
 * Copyright (c) 2005 IIM Lab. All  Rights Reserved.
 * Created by xuehao at Dec 5, 2005
 * Contact: [e-mail]
 * *****************************************************
 */
package org.indigo.db;

import java.io.File;
import java.util.Iterator;
import java.util.Vector;

import org.indigo.xml.XmlTableReader;

/**
 * 采集辅助工具。
 * 读取指定目录中的xml文件。
 * @author wbz
 *
 */
public class Xml2DBTool
{
    private Vector itsFiles=null;
    private String itsRootPath=null;
    public Xml2DBTool()
    {
        
    }
    public void setRootPath( String path )
    {
        itsRootPath = path;
        getFiles( path );
    }
    private void getFiles( String fileName )
    {
        File source = new File( fileName );
        String []files = source.list();
        for( int i=0; i<files.length; i++ )
        {
            File file = new File( source.getPath(), files[i] );
            if( file.isDirectory() )
            {
                getFiles( file.toString() );
            }else
            {
                String str = file.getAbsolutePath();
                if( itsFiles==null )
                {
                    itsFiles = new Vector();
                }
                if( str.endsWith(".xml") )
                    itsFiles.add( str );
//                System.out.println( str );
            }
        }
    }
    public Iterator getFiles()
    {
        return itsFiles.iterator();
    }
    public static void main(String[] args)
    {
        String name;
        Xml2DBTool tool = new Xml2DBTool();
        name = "D:/IExtract/data/农业网站/安徽省";
        tool.setRootPath( name );
        
        Iterator it = tool.getFiles();
        while( it.hasNext() )
        {
            name = (String) it.next();
//            System.out.println( name );
            System.out.println( "created the xml file "+ name );
            XmlTableReader reader = new XmlTableReader( name );
            reader.readAll();
            System.out.println( name + " been have inserted the data into db." );
        }
        
    }
}