每個人都知道我們可以使用DocumentTraversal的NodeIterator遍歷整個xml文檔。 我的應用程序需要一些額外的工作,所以我決定在java stack <>的支持下編寫我自己的XML遍歷程序。java中的XML文檔遍歷器
這裏是我的代碼(我不擅長編碼,所以代碼和邏輯可能看起來很亂)。
public class test
{
private static Stack<Node> gStack = new Stack<Node>();
public static void main(String[] args) throws XPathExpressionException
{
String str =
"<section>"
+ "<paragraph>This example combines regular wysiwyg editing of a document with very controlled editing of semantic rich content. The main content can be"
+ "edited like you would in a normal word processor. Though the difference is that the content remains schema valid XML because Xopus will not allow you to perform actions"
+ "on the document that would render it invalid.</paragraph>"
+ "<paragraph>The table is an example of controlled style. The style of the table is controlled by three attributes:</paragraph>"
+ "<unorderedlist>"
+ "<item><paragraph><emphasis>alternaterowcolor</emphasis>, do all rows have the same color, or should the background color alternate?</paragraph></item>"
+ "<item><paragraph><emphasis>border</emphasis>, a limited choice of border styles.</paragraph></item>"
+ "<item><paragraph><emphasis>color</emphasis>, a limited choice of colors.</paragraph></item>"
+ "</unorderedlist>"
+ "<paragraph>You have quite some freedom to style the table, but you can't break the predefined style.</paragraph>"
+ "</section>";
Document domDoc = null;
try
{
DocumentBuilderFactory docFactory = DocumentBuilderFactory.newInstance();
DocumentBuilder docBuilder = docFactory.newDocumentBuilder();
ByteArrayInputStream bis = new ByteArrayInputStream(str.getBytes());
domDoc = docBuilder.parse(bis);
}
catch (Exception e)
{
e.printStackTrace();
}
Element root = null;
NodeList list = domDoc.getChildNodes();
for (int i = 0; i < list.getLength(); i++)
{
if (list.item(i) instanceof Element)
{
root = (Element) list.item(i);
break;
}
}
NodeList nlist = root.getChildNodes();
System.out.println("root = " + root.getNodeName() + " child count = " + nlist.getLength());
domTraversor(root);
}
private static void domTraversor(Node node)
{
if (node.getNodeName().equals("#text"))
{
System.out.println("textElem = " + node.getTextContent());
if (node.getNextSibling() != null)
{
gStack.push(node.getNextSibling());
domTraversor(node.getNextSibling());
}
else
{
if (node.getParentNode().getNextSibling() != null)
domTraversor(node.getParentNode().getNextSibling());
}
}
else
{
if (node.getChildNodes().getLength() > 1)
{
gStack.push(node);
Node n = node.getFirstChild();
if (n.getNodeName().equals("#text"))
{
System.out.println("textElem = " + n.getTextContent());
if (n.getNextSibling() != null)
{
gStack.push(n.getNextSibling());
domTraversor(n.getNextSibling());
}
}
else
{
gStack.push(n);
domTraversor(n);
}
}
else if (node.getChildNodes().getLength() == 1)
{
Node fnode = node.getFirstChild();
if (fnode.getChildNodes().getLength() > 1)
{
gStack.push(fnode);
domTraversor(fnode);
}
else
{
if (!fnode.getNodeName().equals("#text"))
{
gStack.push(fnode);
domTraversor(fnode);
}
else
{
System.out.println("textElem = " + fnode.getTextContent());
if (fnode.getNodeName().equals("#text"))
{
if (node.getNextSibling() != null)
{
gStack.push(node.getNextSibling());
domTraversor(node.getNextSibling());
}
else
{
if (!gStack.empty())
{
Node sibPn = gStack.pop();
if (sibPn.getNextSibling() == null)
{
sibPn = gStack.pop();
}
domTraversor(sibPn.getNextSibling());
}
}
}
else
{
if (fnode.getNextSibling() != null)
{
domTraversor(fnode.getNextSibling());
}
else
{
if (!gStack.empty())
{
Node sibPn = gStack.pop().getNextSibling();
domTraversor(sibPn);
}
}
}
}
}
}
}
}
}
它和一些xml文檔一起工作正常,但沒有帶有標籤的文檔。
<unorderedlist>
<item>
<paragraph>
<emphasis>alternaterowcolor</emphasis>
, do all rows have the same color, or should the background
color
alternate?
</paragraph>
</item>
<item>
<paragraph>
<emphasis>border</emphasis>
, a limited choice of border styles.
</paragraph>
</item>
<item>
<paragraph>
<emphasis>color</emphasis>
, a limited choice of colors.
</paragraph>
</item>
</unorderedlist>
這裏是如果任何元素有三個以上嵌套子代的場景,我的代碼會停止並且不會繼續。
有沒有人有更好的實施,請提出建議。
不知道你打算做什麼。 XML文檔將被視爲一個樹形結構。要遍歷它們,你需要使用樹遍歷機制,它基本上是一個遞歸函數,像遍歷(節點){對於節點{process(child)中的每個子節點;遍歷(子);}}你的線性traveral不會遍歷樹。 –
爲什麼不使用JAXB解析:http://wiki.processing.org/w/XML_parsing_with_JAXB? – Vikdor
@ d-live你明白了,我需要線性遍歷整個文檔並收集一些元數據,所以我不能使用任何庫。 – Sark