REXML adapter

Purpose

The REXML adapter provides XML processing through Ruby’s standard library, offering maximum portability without requiring any external gems.

Overview

REXML is Ruby’s built-in XML parser, distributed as part of the standard library. It requires no external dependencies and works on all Ruby installations.

Installation

REXML is included with Ruby, so only Moxml needs to be installed:

gem 'moxml'
# No additional gems needed - REXML is in Ruby stdlib

Install:

bundle install

Configuration 

# Explicit selection
context = Moxml.new
context.config.adapter = :rexml

# Global default
Moxml::Config.default_adapter = :rexml

Features

Basic XPath support

REXML provides basic XPath querying:

doc = Moxml.new.parse(xml)

# Basic paths work
books = doc.xpath('//book')
first_book = doc.xpath('/library/book[1]')

# Attribute predicates work
has_id = doc.xpath('//book[@id]')
specific_book = doc.xpath('//book[@id="1"]')

# Position predicates work
first_three = doc.xpath('//book[position() < 4]')

Limited namespace support

REXML can parse and preserve namespaces but cannot use them in XPath queries:

xml = '<library xmlns:dc="http://purl.org/dc/elements/1.1/"><dc:title>Book</dc:title></library>'

doc = Moxml.new.parse(xml)

# Namespace metadata is preserved
element = doc.root.children.first
puts element.namespace  # Works - returns namespace URI

# But namespace-aware XPath does NOT work
doc.xpath('//dc:title', 'dc' => 'http://purl.org/dc/elements/1.1/')
# => Returns empty, cannot resolve namespace prefix

# Workaround: use element name without namespace
doc.xpath('//title')  # Works - finds the element

All basic node types

Support for standard XML node types:

  • Elements with attributes

  • Text nodes

  • CDATA sections

  • Comments

  • Processing instructions

  • DOCTYPE declarations

  • XML declarations

Standard library advantage

  • Always available - no gem dependencies

  • Maximum portability across Ruby versions

  • Simple deployment - no compilation needed

  • Guaranteed compatibility

Limitations

REXML has several limitations compared to other adapters:

XPath limitations:

  • No namespace-aware XPath queries (see example above)

  • Limited function support

  • Some axes not supported

  • Complex predicates may not work

Performance:

  • Slower parsing than C-based libraries

  • Medium serialization speed

  • Higher memory usage than Ox

Namespace XPath workaround:

# Instead of:
doc.xpath('//ns:element', 'ns' => 'http://example.org')

# Use element name matching:
doc.xpath('//element')

# Then filter in Ruby:
elements.select { |e| e.namespace == 'http://example.org' }

Performance characteristics

Based on benchmarks:

Operation Performance Notes

Parse medium XML

~10-20 ips

Pure Ruby parsing

Serialize medium XML

~500-1,000 ips

Medium speed

XPath queries

~5,000-10,000 ips

Limited XPath

Memory usage

Medium

Pure Ruby overhead

Best use cases

Choose REXML when:

  • No external gems can be used (standard library only)

  • Maximum portability is required

  • Small to medium documents

  • Deployment simplicity is critical

  • C extensions cannot be compiled

  • Basic XPath without namespaces is sufficient

Avoid REXML when:

Example usage

Basic operations 

require 'moxml'

# Configure REXML adapter
context = Moxml.new
context.config.adapter = :rexml

xml = '<library><book id="1">Ruby Programming</book></library>'
doc = context.parse(xml)

# Basic XPath works
book = doc.at_xpath('//book[@id="1"]')
puts book.text  # => "Ruby Programming"

# Modify
book.text = 'Advanced Ruby'
book['edition'] = '2nd'

puts doc.to_xml(indent: 2)

XPath queries 

# Supported patterns
books = doc.xpath('//book')
with_id = doc.xpath('//book[@id]')
specific = doc.xpath('//book[@id="1"]')
first_two = doc.xpath('//book[position() <= 2]')

# NOT supported - avoid these
# doc.xpath('//ns:book', namespaces)  # ❌ No namespace XPath
# doc.xpath('count(//book)')          # ⚠️ Limited functions
# doc.xpath('//book[price < 30]')     # ⚠️ May not work

Namespace workarounds 

xml = <<~XML
  <library xmlns:dc="http://purl.org/dc/elements/1.1/">
    <dc:title>Programming</dc:title>
    <dc:creator>Smith</dc:creator>
  </library>
XML

doc = Moxml.new.parse(xml)

# Find elements by name (without namespace prefix)
titles = doc.xpath('//title')

# Check namespace programmatically
titles.each do |title|
  if title.namespace == 'http://purl.org/dc/elements/1.1/'
    puts "DC title: #{title.text}"
  end
end

Comparison with other pure Ruby option

Aspect REXML Oga

XPath support

Limited

Full XPath 1.0

Performance

Medium

Fast

Memory usage

Medium

Medium

Namespace XPath

Not supported

Full support

Standard library

Yes

No (external gem)

Dependencies

None

None (pure Ruby)

References

See also