converttoxml.java~31~

来自「本程序用JAVA编制」· JAVA~31~ 代码 · 共 30 行

import org.xml.sax.*;
import org.w3c.dom.*;
import org.apache.xerces.dom.*;
import org.apache.xerces.parsers.*;
import org.apache.xml.serialize.*;

// Import packages dealing with XSL transformation
import org.apache.xalan.xslt.*;
import org.apache.xalan.xpath.*;

// Import the Tidy package for HTML to XML transformation
import org.w3c.tidy.*;
/**
 * Retrieve an HTML page, convert the source to XML,
 * and write the result to a file.
 */

public class convertToXml {
  public static void main(String args[]) {
    try {
      //Document doc = XMLHelper.tidyHTML("http://weather.yahoo.com/forecast/UKXX0085_c.html?force_units=1");
      //Document doc = StockHelper.tidyHTML("http://stock.sina.com.cn/cgi-bin/stock/quote/quote.cgi?symbol=600002&country=sh&option=1");
      Document doc = StockHelper.tidyHTML("http://news.sina.com.cn/#china");
      XMLHelper.outputXMLToFile(doc,"new-xml/news1.xml");
    }catch(XMLHelperException xmle){
      xmle.printStackTrace();
    }
  }
}