convert XML to JSON

XML Digester

Maps Xml to JavaScript objects while allowing you to influence the conversion. Thereby some unusual Xml documents can be mapped with ease.

This is my first node module and I have never really programmed in JavaScript before. So be warned. If you do not need the features of this converter, you should probably use one of the alternatives

I had an Xml I needed to convert with an unusual charcteristic:


I needed to preserve the order of the 'nodes' ("crossing", "street", etc.):

{ nodes: [ {name:"crossing"}, {name:"street"}, ... ] }

But none of the converters I tried (see below alternatives), allowed me to do that. Some like xml-stream allow to collect nodes of the same name into an array:

   <node kind="crossing"/>
   <node kind="street"/>
   <node kind="cross-walk"/>
   <node kind="street"/>
   <node kind="end-of-town"/>

but I didn't want to create a stylesheet that first transformed these documents.

There are other mappers like xmldom that create a whole Dom tree, but I wanted something simpler and something that could easily be extended by someone else.

var xml_digester = require("xml-digester");
var digester = xml_digester.XmlDigester({});

var xml = "<root>"
  + "<foo>foo1</foo>"
  + "<bar>bar1></bar>"
  + "<foo>foo2</foo>"
  + "</root>"

digester.digest(xml, function(err, result) {
  if (err) { 
  } else {
    // result will be { root: { foo: [ 'foo1', 'foo2' ], bar: 'bar1>' } }

This is the normal mapping behaviour, which all other converters offer. So it should be possible to simply replace your existing mapper with xml-digester.

If you need to influence the mapping you can declare that certain Xml elements should be converted by a special handler. The declaration supports a very minimal subset of XPath:

  • foo matches all 'foo' elements
  • bar/foo matches all 'foo' elements which have a 'bar' element as parent
  • bar//foo matches all 'foo' elements which have a 'bar' element as ancestor
  • /bar//foo matches all 'foo' elements which have a 'bar' element as root element
  • bar/* matches all elements which have a 'bar' element as parent
  • bar/*/foo matches all 'foo' elements which have a 'bar' element as grand parent

There are three predefined handlers.

  1. The SkipElementsHandler (xml_digester.SkipElementsHandler) that just removes elements (and their child elements)
  2. The OrderElementsHandler (xml_digester.OrderedElementsHandler) that is used to preserve the order of elements (see Why ...)
  3. And the DefaultHandler (xml_digester.DefaultHandler), which is used normally but which can be used by another handler as well

To use the handler you have to set it as an option to the digester:

var xml_digester = require("../lib/xml-digester");

var handler = new xml_digester.OrderedElementsHandler("kind");
var options = {
  "handler": [
    { "path": "nodes/*", "handler": handler}
var digester = xml_digester.XmlDigester(options);
var xml = "<nodes>"
        +   "<crossing/><street/><cross-walk/><street/><end-of-town/>"
        + "</nodes>"

digester.digest(xml, function(err, result) {
  if (err) { 
  } else {
    // result will 
    // { nodes:  [ { kind: 'crossing' }, { kind: 'street' }, ... ] }

Since the name of the nodes should preserved, you can define a property-name (in the above example it is 'kind'). The Xml Element name will then be stored in the JavaScript object in that property.

If you do not give a property-name, the '_name' property will be used. But this property is not 'enumerable', i.e. Object.keys(node) will not list the '_name' property.

If there are multiple paths that match a given Xml element, the handler of the first matching path in the 'handler' array will be used.

You can create your own handler, if you understand some of the inner workings of the XmlDigester. For now I would advise you not to do that, because the API might change. But the doucmentation will come as soon as possible, once I feel the API has stabilized enough.


The above are the ones I tried. See


for the full list.

The Xml Digester is based on Issac Schlueter's sax-js.

That library also works in the browser! It would be great if anyone were willing to test Xml Digester in the browser and send me a pull request ...