node-expat
Motivation
You use node.js for speed? You process XML streams? Then you want the fastest XML parser: libexpat!
Speed
A stupid speed test is supplied in bench.js. We measure how many
25-byte elements a SAX parser can process:
- node-xml (pure JavaScript): 23,000 el/s
- libxmljs (libxml2 binding): 77,000 el/s
- node-expat (libexpat binding, this): 113,000 el/s
These numbers were recorded on a Core 2 2400 MHz.
Instructions
Install node-expat:
npm i node-expat
Usage
Important events emitted by a parser:
{ "use strict"; var expat = parser ; parser = "UTF-8"; parser; parser; parser; parser; parser;}; Use test.js for reference.
API
#on('startElement' function (name, attrs) {})#on('endElement' function (name) {})#on('text' function (text) {})#on('processingInstruction', function (target, data) {})#on('comment', function (s) {})#on('xmlDecl', function (version, encoding, standalone) {})#on('startCdata', function () {})#on('startCdata', function () {})#on('endCdata', function () {})#on('entityDecl', function (entityName, isParameterEntity, value, base, systemId, publicId, notationName) {})#on('error', function (e) {})#stop()pauses#resume()resumes
Error handling
We don't emit an error event because libexpat doesn't use a callback
either. Instead, check that parse() returns true. A descriptive
string can be obtained via getError() to provide user feedback.
Alternatively, use the Parser like a node Stream. write() will emit
error events.
Namespace handling
A word about special parsing of xmlns: this is not neccessary in a bare SAX parser like this, given that the DOM replacement you are using (if any) is not relevant to the parser.