You use node.js for speed? You process XML streams? Then you want the fastest XML parser: libexpat!
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.
Install node-expat:
npm i node-expat
Important events emitted by a parser:
- startElement with
name, attrs - endElement with
name - text with
string
There are more. Use test.js for reference.
It's possible to stop and resume the parser from within element handlers using the parsers stop() and resume() methods.
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.
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.