Skip to main content
This is a pre-production deployment of Warehouse. Changes made here affect the production instance of PyPI (
Help us improve Python packaging - Donate today!

Simple and fast Python built-in type XML serialiser

Project Description


Xmlify is simple and fast Python built-in type XML serialiser. Its purpose is to convert nested structures that are composed of types like dict, list, tuple, str, int, etc. It has no dependencies, uses xml.etree.cElementTree and produces XML structure suitable for humans to read.

Use case

I use it for structured logging into MySQL. Later, when I need to investigation an incident, I can further narrow the result set using MySQL XML functions. You can use it for something else.

Result markup

Because the XML tree is constructed only using xml.etree.cElementTree the output is always valid XML. But it is more restrictive to tag names. Output tag names avoid colon (XML namespaces) and are ASCII-only, even though the specification permits Unicode codepoints. It was to simplify things, as there was no benefit of having Unicode tags, and to maximise XML parser compatibility.

Unicode tag names are hex-encoded and prefixed with x. XML-incompatible binary values are hex-encoded in the same way. XML-incompatible ASCII characters in tag names are replaced with underscore. If tag name starts with a digit it is prefixed with n. Though, it’s easy to override. The following regular expressions control substitution and tag’s first character match. You can also monkeypatch xmlify._key completely.

xmlify._notFirstCharRe  = re.compile(r'[^a-z_]{1}', re.IGNORECASE)
xmlify._notOtherCharsRe = re.compile(r'[^a-z0-9_\-\.]', re.IGNORECASE)

The type information isn’t preserved intentionally to make the output easier to read to a human. If you need to preserve type information or to convert XML to objects two-way, just use stdlib’s xmlrpclib.dumps/xmlrpclib.loads.


This is the public API:

def dump(obj, fp, root = 'data', declaration = False): pass

def dumps(obj, root = 'data', declaration = False): pass

Use it like:

import datetime
import xmlify

d = {
  'python' : {
    2 : {
      2.7 : {
        'version' : (2, 7, 10),
        'date'    :, 5, 23)
    3 : {
      3.3 : {
        'version' : (3, 3, 6),
        'date'    :, 10, 12)
      3.4 : {
        'version' : (3, 4, 3),
        'date'    :, 2, 25)

It prints the following XML (indented separately):



It’s worth just 14 LLOC of a recursive function. The rest ~100 LLOC is supporting code that goes in line with Pareto principle.


$ python -c 'import os; print(len(os.environ))'
$ python2.7 -m timeit 'import os,xmlify; xmlify.dumps(os.environ)'
1000 loops, best of 3: 987 usec per loop
$ python3.3 -m timeit 'import os,xmlify; xmlify.dumps(os.environ)'
1000 loops, best of 3: 1.62 msec per loop
$ pypy -m timeit 'import os,xmlify; xmlify.dumps(os.environ)'
1000 loops, best of 3: 193 usec per loop

Inventing own wheel

NIH was not the case – even though I already had a working code I would happily have used an existing library that fits my needs. At the Cheese Shop there were several groups of libraries that do the same or closely related thing:

  • Mappers that need schema up-front
  • Libraries that need to build dependencies with OS package dependencies, e.g. lxml
  • Marshallers that try to preserve type information, thus making result markup hard to read
  • Libraries that build XML tree manually with strings and thus with potential escaping issues
  • Just broken

I the end I just decided to package the code I had.

Release History

Release History

This version
History Node


History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
Xmlify-0.1.1.tar.gz (5.9 kB) Copy SHA256 Checksum SHA256 Source Jun 9, 2015

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting