C++ Blog - Using XML in C++

Working with Xerces DOM nodes, I had the problem to query DOMNodes with X-Path. Because Xerces doesn't implement any X-Path, I tried the Pathan library. After stumpling around, I finaly tried the Xalan, to work with the Xerces DOMNodes. This is not a simple task after all. Following code should give you an idea how to do this:

//m_pNode is an instance of XERCES_CPP_NAMESPACE::DOMNode 
//strQuery is the X-Path

const XALAN_CPP_NAMESPACE::XalanDOMString 
  strExpression(strQuery.c_str(NULL)); //UNICODE PROBLEM

XERCES_CPP_NAMESPACE::DOMDocument* pDocument = 
  m_pNode->getOwnerDocument(); 

XALAN_CPP_NAMESPACE::XercesParserLiaison mParserLiaison;
mParserLiaison.setBuildWrapperNodes(true);
mParserLiaison.setBuildMaps(true);

XALAN_CPP_NAMESPACE::XalanDocument* pXalanDocument = 
  mParserLiaison.createDocument(pDocument, false, false);

XALAN_CPP_NAMESPACE::XalanElement* pNameSpaceNode = 
  pXalanDocument->getDocumentElement();

XALAN_CPP_NAMESPACE::XercesDocumentWrapper* pXercesDocumentWrapper =
  mParserLiaison.mapDocumentToWrapper(pXalanDocument);

XALAN_CPP_NAMESPACE::XalanNode* pRootContextNode = 
  pXercesDocumentWrapper->mapNode(m_pNode);

XALAN_CPP_NAMESPACE::XercesDOMSupport mDOMSupport;

if (pRootContextNode != NULL)
{
  XALAN_CPP_NAMESPACE::XPathEvaluator mEvaluator;
  XALAN_CPP_NAMESPACE::XalanNode* pFoundXalanNode = 
    mEvaluator.selectSingleNode(mDOMSupport,
                                pRootContextNode,
                                strExpression.c_str(),
                                pNameSpaceNode);

  XERCES_CPP_NAMESPACE::DOMNode* pTargetNode = NULL;
  if (pFoundXalanNode != NULL)
  {
    DOMNode const* pConstTargetNode = pXercesDocumentWrapper->mapNode(pFoundXalanNode);
    pTargetNode = const_cast<DOMNode *>(pConstTargetNode);
    ...
  }
}

Following pages helped out and have further informations:

• The pathan library is a good alternative to Xalan
  http://software.decisionsoft.com/pathanIntro.html
• The API documentation of Xerces C++
  http://xml.apache.org/xerces-c/apiDocs/index.html
• The API documentation of Xalan C++
  http://xml.apache.org/xalan-c/apiDocs/index.html

The MSXML Parser from Microsoft has one big difference to the XML Paser from Apache. Quering a XML node by XPath using the index-operator is different. MSXML starts at 0 (zero), the XML Parser of Xerces starts at 1 (one).

Be careful porting software from one parser to the other parser. The x-path ./MainNode/Node[1] returns on MSXML parser the 2nd child node named node.

To solve this problem I read, that the E5 and later has implemented that [0] should be the first node, but according to the W3C standard it should have been [1]!!

To fix this problem we have to set the SelectionLanguage property of the document to XPath.

Following pages helped out and have further informations:

• How to fix the problem with IE5+
  http://www.w3schools.com/xpath/xpath_examples.asp
• Good reference of the XPath syntax
  http://www.w3schools.com/xpath/xpath_syntax.asp