Binary Data in XML

owain · November 11th, 2003, 10:24 AM

Is there any way of including raw binary data in an XML file?

I have looked at the CDATA section and was under the impression that this could be used to embed raw binary data (including characters that outside of the CDATA section would be seen as illegal characters, not just < and >).

However when I use the MSXML 4.0 component in VB with raw binary data embedded in a CDATA section an error is raised when loading the document. The Microsoft XMLDOM ActiveX component in a web page on the other hand has no problems loading and parsing the embedded raw binary data, nor does IE6 for that matter.

Can anyone tell me the definitive answer to whether raw binary data can be included in a CDATA section of an XML file?

Regards
Owain Williams

joefawcett · November 11th, 2003, 11:22 AM

In xml 1.0 you cannot have any character belo 0x20 except for 9 (tab), (10) newline and 13 (carriage return.
To embed others you can use (amongst other things) Base64 encoding. It's very easy if you are using msxml parser and here are two scripts to try it out, create a simple MS Word doc called hello1.doc or another file if you wish but change line two as appropriate:

Code:

//Save this as BinaryToXml.js
var WORD_PATH = "hello1.doc";
var XML_PATH = "word.xml"

function main()
{
  var oDoc = new ActiveXObject("MSXML2.DOMDocument.4.0");
  oDoc.async = false;
  var oStream = new ActiveXObject("ADODB.Stream");
  oStream.type = 1; //Binary
  oStream.open();
  oStream.loadFromFile(WORD_PATH);
  var oNode = oDoc.createProcessingInstruction("xml", "version='1.0' encoding='utf-8'")
  oDoc.insertBefore(oNode, oDoc.childNodes[0]);
  var oRoot = oDoc.createElement("Root")
  oDoc.documentElement = oRoot
  oRoot.setAttribute("xmlns:dt", "urn:schemas-microsoft-com:datatypes");
  oNode = oRoot.appendChild(oDoc.createElement("wordDoc"));
  oNode.dataType = "bin.base64";
  oNode.nodeTypedValue = oStream.read();
  oStream.close();
  oDoc.save(XML_PATH);
}

main();

Then this to read from first file:

Code:

//Save this as XmlToBinary.js
var WORD_PATH = "hello2.doc";
var XML_PATH = "word.xml"

function main()
{
  var oDoc = new ActiveXObject("MSXML2.DOMDocument.4.0");
  oDoc.async = false;
  oDoc.load(XML_PATH);
  var oStream = new ActiveXObject("ADODB.Stream");
  oStream.type = 1; //Binary
  oStream.open();
  oStream.write(oDoc.documentElement.selectSingleNode("wordDoc").nodeTypedValue);
  oStream.savetoFile(WORD_PATH);
  oStream.close();  
}

main();

Joe (MVP - xml)