dom4j如何获取节点的行数，列数

在使用dom4j解析xml文件时，可能会对一些节点做检测，判断是否符合schema，对一些不符合的节点要作出提示。为了使作出的提示更友好，还需要指出错误在哪里。但是dom4j并没有提供相关的功能，或者说这个功能隐藏的很深。搜索了一下，发现这个问题的答案很少。我在一个mail（https://www.mail-archive.com/[email protected]/msg02769.html）中得到了提示，摸索了出来，巨麻烦。该mail提到了主要的过程：

org.xml.sax.Locator locator = new …;
DocumentFactory documentFactory = new DocumentFactoryWithLocator(locator);
SAXContentHandler contentHandler = new SAXContentHandler(documentFactory);
contentHandler.setDocumentLocator(locator);
org.xml.sax.XMLReader reader = …;
reader.setContentHandler(contentHandler );
reader.parse(…);
Document document = contentHandler.getDocument();

在DocumentFactory中：

public Element createElement(QName qname) {
	ElementWithLocation element =  new ElementWithLocation (qname);
    element.setLocation(locator.getLineNumber(),
	locator.getColumnNumber());
    return element;
}

方向是对的，但是按照这个方法做出来，却得不到正确的结果。原因是，SAXReader中的XMLReader在调用parse方法时，会重新赋值locator，把我传进去的locator覆盖掉了，这样在DocumentFactory就不是XMLReader中的locator。dom4j使用的XMLReader是org.apache.xerces中的SAXParser，该类继承了AbstractSAXParser，在AbstractSAXParser中有一个方法：

public void startDocument(XMLLocator locator, String encoding, 
                              NamespaceContext namespaceContext, Augmentations augs)
        throws XNIException {
        
        fNamespaceContext = namespaceContext;

        try {
            // SAX1
            if (fDocumentHandler != null) {
                if (locator != null) {
                    fDocumentHandler.setDocumentLocator(new LocatorProxy(locator));
                }
                fDocumentHandler.startDocument();
            }

            // SAX2
            if (fContentHandler != null) {
                if (locator != null) {
                    fContentHandler.setDocumentLocator(new LocatorProxy(locator));
                }
                fContentHandler.startDocument();
            }
        }
        catch (SAXException e) {
            throw new XNIException(e);
        }

    }

就是这个方法覆盖了contentHandler的locator。我的做法就是在AbstractSAXParser重新赋值locator的时候获取这个值，传递给DocumentFactory。

上代码，首先需要扩展原来的Element类，使之可以记录节点的位置信息（需要记录Attribute等的同理）。

public class GokuElement extends DefaultElement {
	private int lineNum = 0, colNum = 0;

	public GokuElement(QName qname) {
		super(qname);
		// TODO Auto-generated constructor stub
	}

	public GokuElement(QName qname, int attrCount) {
		super(qname, attrCount);
	}

	public GokuElement(String name) {
		super(name);
	}

	public GokuElement(String name, Namespace namespace) {
		super(name, namespace);
	}

	public int getColumnNumber() {
		return this.colNum;
	}

	public int getLineNumber() {
		return this.lineNum;
	}

	public void setLocation(int lineNum, int colNum) {
		this.lineNum = lineNum;
		this.colNum = colNum;
	}
}

然后，扩展DocumentFactory，让factory生成我们定义的Element：

public class DocumentFactoryWithLocator extends DocumentFactory {
	private Locator locator;

	public DocumentFactoryWithLocator(Locator locator) {
		super();
		this.locator = locator;
	}

	@Override
	public Element createElement(QName qname) {
		GokuElement element = new GokuElement(qname);
		element.setLocation(this.locator.getLineNumber(), this.locator.getColumnNumber());
		return element;
	}

	@Override
	public Element createElement(String name) {
		GokuElement element = new GokuElement(name);
		element.setLocation(this.locator.getLineNumber(), this.locator.getColumnNumber());
		return element;
	}

	public void setLocator(Locator locator) {
		this.locator = locator;
	}
}

然后扩展SAXContentHandler：

public class GokuSAXContentHandler extends SAXContentHandler {
	private DocumentFactoryWithLocator documentFactory = null;

	public GokuSAXContentHandler(DocumentFactory documentFactory2, ElementHandler dispatchHandler) {
		// TODO Auto-generated constructor stub
		super(documentFactory2, dispatchHandler);
	}

	public void setDocFactory(DocumentFactoryWithLocator fac) {
		this.documentFactory = fac;
	}

	@Override
	public void setDocumentLocator(Locator documentLocator) {
		super.setDocumentLocator(documentLocator);
		if (this.documentFactory != null)
			this.documentFactory.setLocator(documentLocator);
	}
}

最后扩展SAXReader

public class GokuSAXReader extends SAXReader {
	DocumentFactory docFactory;
	Locator locator;

	public GokuSAXReader(DocumentFactory docFactory) {
		// TODO Auto-generated constructor stub
		super(docFactory);
		this.docFactory = docFactory;
	}

	public GokuSAXReader(DocumentFactory docFactory, Locator locator) {
		// TODO Auto-generated constructor stub
		super(docFactory);
		this.locator = locator;
		this.docFactory = docFactory;
	}

	@Override
	protected SAXContentHandler createContentHandler(XMLReader reader) {
		return new GokuSAXContentHandler(this.getDocumentFactory(), super.getDispatchHandler());
	}

	@Override
	public Document read(InputSource in) throws DocumentException {
		try {
			XMLReader reader = this.getXMLReader();

			reader = this.installXMLFilter(reader);

			EntityResolver thatEntityResolver = super.getEntityResolver();

			if (thatEntityResolver == null) {
				thatEntityResolver = this.createDefaultEntityResolver(in.getSystemId());
				super.setEntityResolver(thatEntityResolver);
			}

			reader.setEntityResolver(thatEntityResolver);

			SAXContentHandler contentHandler = this.createContentHandler(reader);
			contentHandler.setEntityResolver(thatEntityResolver);
			contentHandler.setInputSource(in);

			boolean internal = this.isIncludeInternalDTDDeclarations();
			boolean external = this.isIncludeExternalDTDDeclarations();

			contentHandler.setIncludeInternalDTDDeclarations(internal);
			contentHandler.setIncludeExternalDTDDeclarations(external);
			contentHandler.setMergeAdjacentText(this.isMergeAdjacentText());
			contentHandler.setStripWhitespaceText(this.isStripWhitespaceText());
			contentHandler.setIgnoreComments(this.isIgnoreComments());
			reader.setContentHandler(contentHandler);

			this.configureReader(reader, contentHandler);
			((GokuSAXContentHandler) contentHandler).setDocFactory((DocumentFactoryWithLocator) this.docFactory);
			contentHandler.setDocumentLocator(this.locator);
			reader.parse(in);

			return contentHandler.getDocument();
		} catch (Exception e) {
			if (e instanceof SAXParseException) {
				// e.printStackTrace();
				SAXParseException parseException = (SAXParseException) e;
				String systemId = parseException.getSystemId();

				if (systemId == null) {
					systemId = "";
				}

				String message = "Error on line " + parseException.getLineNumber() + " of document " + systemId + " : "
						+ parseException.getMessage();

				throw new DocumentException(message, e);
			} else {
				throw new DocumentException(e.getMessage(), e);
			}
		}
	}
}

重写了read（InputSource）方法，使用我们写的SAXContentHandler。其他签名的read方法最后都调用了这个read方法。

解析文件的代码：

Locator locator = new LocatorImpl();
DocumentFactory docFactory = new DocumentFactoryWithLocator(locator);
SAXReader reader = new GokuSAXReader(docFactory, locator);
Document doc = reader.read(new File("goku.xml"));

需要获取Element信息时：

System.out.println(((GokuElement) element).getLineNumber());