使用xpath获取子节点？

Question

I have the following xml. I need to get all the child nodes of root node using an xpath query. how can i write the xpath expression?

<rootElement> 

  <rootElementOne xmlns="http://some.com"> 
    <rootElementTwo> 
      <Id>12345</balId> 
      <name>Name1</businessName> 
     </rootElementTwo> 
  </rootElementOne> 

  <rootElementOne xmlns="http://some.com"> 
    <rootElementTwo> 
      <Id>6789</balId> 
      <name>Name2</businessName> 
     </rootElementTwo> 
  </rootElementOne>  

</rootElement>

The expression should return the result below:

      <rootElementOne xmlns="http://some.com"> 
        <rootElementTwo> 
          <Id>12345</balId> 
          <name>Name1</businessName> 
         </rootElementTwo> 
      </rootElementOne> 

      <rootElementOne xmlns="http://some.com"> 
        <rootElementTwo> 
          <Id>6789</balId> 
          <name>Name2</businessName> 
         </rootElementTwo> 
      </rootElementOne>

I tried using rootElement/rootElementOne/* but no result.

Thanks!

Answer 1

Watch out for terminology here. In XML, at least in XPath terminology, the "root node" is the (invisible) ancestor of all elements, text nodes, comments, processing instructions, and other nodes in a document. The root node is addressed by the XPath expression / . It is not an element, but is the parent of the outermost element , aka the document element . In your XML document, the root node is the parent of <rootElement> .

All "child nodes of the root node" would be selected by this XPath expression:

/node()

but that would return one element, namely <rootElement> , which is not the result you want.

Instead, you probably want all child nodes of the document element , so this is your XPath expression:

/*/node()

This will return the <rootElementOne> elements, and (depending on your settings) also the text node between them, which consists of whitespace.

Alternatively, maybe you want all element children of the document element. In other words, you don't care about text nodes, comments, or anything besides elements. (A lot of people who are unfamiliar with the details of XML say "node" when they mean "element node".)

If that's what you want, the XPath expression for it is

/*/*

or in your case, you could do

/rootElement/some:rootElementOne

where some is declared outside the XPath as a namespace prefix for http://some.com . If you want to know how to declare a namespace prefix for XPath in Java, let us know, and show us what Java code you're already using to call XPath. Or better yet, search on this site because there are already good answers with example code.

When you tried rootElement/rootElementOne/* , you selected nothing, because of namespaces. An XPath step of the form rootElementOne (in XPath 1.0) means "an element named rootElementOne in no namespace." (In XPath 2.0, it means "in the default XPath namespace," and there are ways outside of XPath to set the default XPath namespace.) So you asked for rootElementOne in no namespace, whereas your <rootElementOne> elements are in the http://some.com namespace.

If you want to be namespace-agnostic, you can use * instead of rootElementOne , or you could use *[local-name() = 'rootElementOne'] . However, if you do this because you don't know how to use namespaces in XML and XPath, they will probably continue to be a thorn in your flesh until you learn. :-)

Once that is fixed, you should get two <some:rootElementTwo> elements (because you asked for child elements of rootElementOne ), but this would only work in the context of the root node of the document. That's because an XPath expression that starts with an element name X is really starting with child::X , meaning the child of the context node. If you don't know what the context node is at the time, or don't want to be dependent on it, start your XPath expression with / or // . That tells XPath to start from the root node of the document.

Answer 2

/rootElement/* will return all rootElementOne where you can do some action on

/rootElement//* will return everything below rootElement