# Xml parser + dom + robust + speed =?

**URL:** <https://www.ruby-forum.com/t/xml-parser-dom-robust-speed/97671>\
**Category:** Ruby\
**Created:** [July 9, 2006, 11:54am UTC](https://www.ruby-forum.com/t/xml-parser-dom-robust-speed/97671 "2006-07-09T11:54:31Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![pete](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/3ab097/32.png) [@pete](https://www.ruby-forum.com/u/pete)\
**Post date:** [July 9, 2006, 11:54am UTC](https://www.ruby-forum.com/t/xml-parser-dom-robust-speed/97671/1 "2006-07-09T11:54:31Z")

</div>

Hi,

can somebody suggest a good parser that offers:

- multiple encodings
- production-like quality
- dom tree manipulation
- good performance

is REXML the only choice or are there others?

Cheers  
Peter

---

<div class="post-metadata">

**Author:** ![pete](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/3ab097/32.png) [@pete](https://www.ruby-forum.com/u/pete)\
**Post date:** [July 9, 2006, 1:32pm UTC](https://www.ruby-forum.com/t/xml-parser-dom-robust-speed/97671/2 "2006-07-09T13:32:06Z")

</div>

Pete wrote:

> Hi,
> 
> can somebody suggest a good parser that offers:
> 
> - multiple encodings
> - production-like quality
> - dom tree manipulation
> - good performance
> 
> is REXML the only choice or are there others?

There’s also libxml and probably others:  
[http://www.google.com/search?q=ruby+xml](http://www.google.com/search?q=ruby+xml)

However, note that DOM and performance nearly automatically exclude each  
other. Depending on what you do stream parsing is almost always faster  
and uses less memory.

Kind regards

```
robert

```

---

<div class="post-metadata">

**Author:** ![pete](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/p/3ab097/32.png) [@pete](https://www.ruby-forum.com/u/pete)\
**Post date:** [July 10, 2006, 6:01pm UTC](https://www.ruby-forum.com/t/xml-parser-dom-robust-speed/97671/3 "2006-07-10T18:01:07Z")

</div>

If you are working under windows, it seems to be difficult to get some  
of the other XML libraries to compile. Like the previous poster said,  
libXml should be pretty speedy. Rexml is the only pure ruby library  
though (there are some built on top of rexml).

Try parsing a document, and see how the performance turns out, if it  
really is too slow for you try the following things:

1. Avoid using the xpath lookups if you don’t need them. They’re the  
most natural way to access xml data in my opinion, but they are a bit  
slower
2. Start using the stream parser and your DOM tree yourself with  
lighterweight nodes, if you need an examle let me know, I can post one.

Enjoy,  
.adam sanderson