【Android】SAX解析之错误纠正

               

在讲这次错误之前，先看一下下面这段代码。  【◆以下解析方法是错误的×】

import java.util.ArrayList;import java.util.List;import org.xml.sax.Attributes;import org.xml.sax.SAXException;import org.xml.sax.helpers.DefaultHandler;import android.util.Log;public class XmlHandler extends DefaultHandler{  private final String TAG = this.getClass().getSimpleName();  /**XML文件中标签定义*/ private final String TAG_Article = "Article"; private final String TAG_ArticleID = "ArticleID"; private final String TAG_Title = "Title"; private final String TAG_Date = "Date"; private final String TAG_SmallPictures = "SmallPictures"; private final String TAG_LargePictures = "LargePictures"; private final String TAG_Category = "Category"; private static final String TAG_HeadNote = "HeadNote"; private static final String TAG_SubTitle = "SubTitle"; private static final String TAG_Source = "Source";  //当前正在解析的TAG private String currentName;  //单个文章 private News news = null;  //文章列表 private List  newsList = null;  //解析开始时间 private long start_time;  private boolean flag = false;  @Override public void characters(char[] ch, int start, int length)   throws SAXException {  super.characters(ch, start, length);    if(!flag) {   return;  }  // 取值  String value = new String(ch, start, length);  Log.d(TAG, "Element: " + currentName  + " Element Value: " + value);  if(value != null) {   if(TAG_ArticleID.equals(currentName)) {    news.setArticleId(value);   } else if(TAG_Title.equals(currentName)) {    news.setTitle(value);   } else if(TAG_Date.equals(currentName)) {    news.setDate(value);   } else if(TAG_Category.equals(currentName)) {    news.setCategory(value);   } else if(TAG_SmallPictures.equals(currentName)) {    news.setSmallPicture(value);   } else if(TAG_LargePictures.equals(currentName)) {    news.setLargePicture(value);   } else if(TAG_HeadNote.equals(currentName)) {    news.setHeadNote(value);   } else if(TAG_SubTitle.equals(currentName)) {    news.setSubTitle(value);   } else if(TAG_Source.equals(currentName)) {    news.setSource(value);   }  } } @Override public void startDocument() throws SAXException {  super.startDocument();    start_time = System.currentTimeMillis();  newsList = new ArrayList(); } @Override public void startElement(String uri, String localName, String qName, Attributes attributes) throws SAXException {  super.startElement(uri, localName, qName, attributes);  this.currentName = localName;  flag = true;  if(TAG_Article.equals(localName)) {   news = new News();  } }  @Override public void endElement(String uri, String localName, String qName)   throws SAXException {  super.endElement(uri, localName, qName);  flag = false;    if(TAG_Article.equals(localName)) {   newsList.add(news);  } }  @Override public void endDocument() throws SAXException {  super.endDocument();    long end = System.currentTimeMillis();  Log.d(TAG, "Parse List's Xml Cost: " + (end - start_time) + " !!"); }}

Baidu 或者 Google 一下 “Android Sax 解析” ， 给出的Sample无一例外都是如此。 坑爹啊... 甚至连有些书籍中都是这么写的， 比如《Android开发入门与实践》。（本书亲自确认过，其他书情况不详）

没错， 一般情况下，这么写是可以的， 而且在大多数情况下解析出来也是正确的。 但是就是偶尔会出错， 这个时候通常你都莫不着头脑， 怎么回事? 数据没错啊，解析部分代码貌似也没问题.. 真是奇了怪了。 其实问题都出在上面那段代码上！！

大家都认为 SAX 解析过程大致如下：

 startDocument  ->   startElement  -> characters -> endElement -> endDocument

没错，就是这样， startElement  读取起始标签， endElement 读取结束标签，characters 呢？当然是读取其值， 这没错，但是大家都天真的以为 characters 只执行一次，并且一次就读取了全部内容。错就错在这！

其实characters 是很有可能会执行多次的，当遇到内容中有回车，\t等等内容时，它很有可能就执行多次。 有的人可能会说，那我没有这些是不是就只执行一次了？ 看下我实测结果：

 

测试用XML如下：

<News> <Article>  <ArticleID>1000555ArticleID>  <Title>Title>  <Date>2011-11-25 14:23:52Date>  <SmallPictures>livenews/images/s20.pngSmallPictures>  <LargePictures>livenews/images/l20.pngLargePictures>  <Category>闻天下Category>  <HeadNote>HeadNote>  <SubTitle>SubTitle>  <Author>Author>  <Source>人民日报Source>  <Abstract>Abstract> Article> <Article>  <ArticleID>1000554ArticleID>  <Title>Title>  <Date>2011-11-25 14:22:33Date>  <Category>Category>  <HeadNote>HeadNote>  <SubTitle>SubTitle>  <Author>Author>  <Source>Source>  <Abstract>Abstract> Article> <Article>  <ArticleID>1000553ArticleID>  <Title>Title>  <Date>2011-11-25 14:21:23Date>  <SmallPictures>livenews/images/s21.pngSmallPictures>  <LargePictures>livenews/images/l21.pngLargePictures>  <Category>Category>  <HeadNote>HeadNote>  <SubTitle>SubTitle>  <Author>Author>  <Source>Source>  <Abstract>Abstract> Article><News>

可以很明显的看到，在解析 1000553  这一段时， characters执行了两次，将内容"1000553"分两次读取.. 用上面那种方式的最终结果就是 ArticleID = 00553 了。 那如果你的应用需要根据这个id 进一步获取内容岂不是死翘翘了？（比如这边根据id获取新闻详细内容）

好了，废话不多说了，看下正确的写法！  【★以下解析方法才是正确的 √ 】

import java.util.ArrayList;import java.util.List;import org.xml.sax.Attributes;import org.xml.sax.SAXException;import org.xml.sax.helpers.DefaultHandler;import android.util.Log;public class XmlHandler extends DefaultHandler{  private final String TAG = this.getClass().getSimpleName();  /**XML文件中标签定义*/ private final String TAG_Article = "Article"; private final String TAG_ArticleID = "ArticleID"; private final String TAG_Title = "Title"; private final String TAG_Date = "Date"; private final String TAG_SmallPictures = "SmallPictures"; private final String TAG_LargePictures = "LargePictures"; private final String TAG_Category = "Category"; private static final String TAG_HeadNote = "HeadNote"; private static final String TAG_SubTitle = "SubTitle"; private static final String TAG_Source = "Source";  //单个文章 private News news = null;  //文章列表 private List  newsList = null;  //解析开始时间 private long start_time;  //(1) private StringBuilder sb = new StringBuilder();  @Override public void characters(char[] ch, int start, int length)   throws SAXException {  super.characters(ch, start, length);    //(2)不管在startElement到endElement的过程中，执行了多少次characters， 都会将内容添加到StringBuilder中，不会丢失内容  sb.append(ch, start, length); } @Override public void startDocument() throws SAXException {  super.startDocument();    start_time = System.currentTimeMillis();  newsList = new ArrayList(); } @Override public void startElement(String uri, String localName, String qName, Attributes attributes) throws SAXException {  super.startElement(uri, localName, qName, attributes);  //(3) 开始收集新的标签的数据时，先清空历史数据  sb.setLength(0);  if(TAG_Article.equals(localName)) {   news = new News();  } }  @Override public void endElement(String uri, String localName, String qName)   throws SAXException {  super.endElement(uri, localName, qName);    //(4)原来在characters中取值，现改在此取值  String value = sb.toString();    if(TAG_ArticleID.equals(localName)) {   news.setArticleId(value);  } else if(TAG_Title.equals(localName)) {   news.setTitle(value);  } else if(TAG_Date.equals(localName)) {   news.setDate(value);  } else if(TAG_Category.equals(localName)) {   news.setCategory(value);  } else if(TAG_SmallPictures.equals(localName)) {   news.setSmallPicture(value);  } else if(TAG_LargePictures.equals(localName)) {   news.setLargePicture(value);  } else if(TAG_HeadNote.equals(localName)) {   news.setHeadNote(value);  } else if(TAG_SubTitle.equals(localName)) {   news.setSubTitle(value);  } else if(TAG_Source.equals(localName)) {   news.setSource(value);  }    if(TAG_Article.equals(localName)) {   newsList.add(news);  } }  @Override public void endDocument() throws SAXException {  super.endDocument();    long end = System.currentTimeMillis();  Log.d(TAG, "Parse List's Xml Cost: " + (end - start_time) + " !!"); }}

归纳为三点：

1.startElement的时候， new StringBuilder(); 或者 sb.setLength(0); (我建议后者)
2.characters的时候，sb.append(ch, start, length);
3.endElement的时候，sb.toString(); 此时StringBuilder中的内容才是解析的结果

通过这种方法就不会再有数据离奇丢失的情况了(同时也不需要像错误方法那样再设个currentTag之类的了，逻辑繁杂了，还出错)！ 

希望大家可以尽早看到这篇文章，不要继续被吭了！！！