为什么StAX创建XML比DOM慢？

Question

我试图测量StAX和DOM创建两个相同文档所需的时间。 我不知道为什么DOM在创建XML方面更快。 也许我的StAX编写器代码不是很好。 所以这里是StAX（更长的代码）

  public static final int pocet =100000;

try {

  String encoding = "UTF-8";
  XMLOutputFactory f = XMLOutputFactory.newInstance();
  XMLStreamWriter w = f.createXMLStreamWriter(
                      new FileOutputStream(subor),
                      encoding);
  w.writeStartDocument(encoding, "1.0");
  w.writeCharacters("\r\n");
  w.writeStartElement("Noviny");        
  for (int i = 1;  i <= pocet;  i++) {
    w.writeCharacters("\r\n  ");
    w.writeStartElement("Autor");
    w.writeCharacters("\r\n  ");
    w.writeStartElement("Id"); 
    String ID = Integer.toString(i);
    w.writeCharacters(ID);
    w.writeEndElement();
    w.writeCharacters("\r\n  ");
    w.writeStartElement("Meno");
    w.writeCharacters("Autor"+i);
    w.writeEndElement();
    w.writeCharacters("\r\n  ");
    w.writeStartElement("Email");
    w.writeCharacters("Autor"+i+"@email.com");
    w.writeEndElement();
    w.writeCharacters("\r\n  ");
    w.writeStartElement("Tel_cislo");
    w.writeAttribute("typ", "pevna");
    w.writeCharacters("+4219");
    w.writeEndElement();
    w.writeCharacters("\r\n  ");
    w.writeStartElement("plat");

    w.writeCharacters("5000");
    w.writeEndElement();
    w.writeCharacters("\r\n  ");        
    w.writeEndElement(); 
    w.writeCharacters("\r\n");



  }
  w.writeCharacters("\r\n");
  w.writeEndElement();        
  w.writeCharacters("\r\n");
  w.writeEndDocument();
  w.close(); 



}
catch (Exception e) {
  e.printStackTrace();
}

DOM

       int pocet =1500000;
try {
 DocumentBuilderFactory dfactory = DocumentBuilderFactory.newInstance();
 DocumentBuilder builder = dfactory.newDocumentBuilder();
 DOMImplementation domImpl = builder.getDOMImplementation();
 Document document = domImpl.createDocument(null, "Noviny", null);
 Node noviny = document.getDocumentElement();
 for(int i = 0; i<pocet ; i++) {
 Element autor = document.createElement("autor");
 Element meno = document.createElement("meno");
 Element id = document.createElement("id");
 Element email = document.createElement("email");
 Element tel = document.createElement("tel_cislo");
  Element plat = document.createElement("plat");
 Text textid = document.createTextNode(""+i);
 Text textmeno = document.createTextNode("Autor"+i);
 Text textemail = document.createTextNode("mail@gmail.com");
 Text texttel = document.createTextNode("65456465465");
 Text textplat = document.createTextNode("200");
   noviny.appendChild(autor);
    autor.appendChild(id);
        id.appendChild(textid);
    autor.appendChild(meno);
        meno.appendChild(textmeno);
    autor.appendChild(email);
        email.appendChild(textemail);
    autor.appendChild(tel);
        tel.appendChild(texttel);
    autor.appendChild(plat);
        plat.appendChild(textplat);
     }  
  TransformerFactory factory = TransformerFactory.newInstance();
  Transformer transformer = factory.newTransformer();
  transformer.setOutputProperty(OutputKeys.INDENT, "yes"); 
  transformer.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
  transformer.transform(new DOMSource(document),new StreamResult(new File(subor)));

我知道StAX在阅读文档方面更快，但我无法解释这一点。 除了我有很糟糕的StAX代码

谢谢

Answer 1

在这里添加BufferedOuputStream

XMLStreamWriter w = f.createXMLStreamWriter(new BufferedOutputStream(new FileOutputStream(subor)), "UTF-8");

现在测试速度，你会发现StAX的速度至少要快2倍