Python ships with XML libraries, but none addressing XML Schema validation. Several options are available to make up for this feature. The solution demonstrated in this Python Tip is a basic implementation of XML schema validation using MSXML and PyWin32.

Related article: XML Validation with Python, MSXML and comtypes


In order to try the code and examples exposed in this article, it is assumed that you have Python already installed on your system and that you are familiar with this popular programming language. In addition you will need to the following software:


  • ActivePython is a Python distribution for Windows shipping with PyWin32, therefore you don't need to install PyWin32 if you have installed Python from ActiveState.
  • For additional resources, see section Resources at the end of this post.


The code in this section demonstrates a basic idea to validate an XML document with PyWin32 and MSXML. The XML files and the XML Schema files used are respectively:

books.xsd (XML Schema)

<?xml version="1.0" encoding="UTF-8"?>
<!-- $Id: books.xsd 435 2008-01-20 02:15:28Z andre $ -->
<xsd:schema xmlns=""

  <xsd:element name="Books">
        <xsd:element ref="Book" maxOccurs="unbounded"/>

  <xsd:element name="Book">
        <xsd:element ref="Title"/>
        <xsd:element ref="Authors"/>
      <xsd:attribute name="isbn" use="required">
          <xsd:restriction base="xsd:string">
            <xsd:pattern value="[0-9]{10}"/>

  <xsd:element name="Authors">
        <xsd:element ref="Author" maxOccurs="unbounded"/>

  <xsd:element name="Title" type="xsd:string" />
  <xsd:element name="Author" type="xsd:string" />


books.xml (Valid XML)

<?xml version="1.0" encoding="UTF-8"?>
<!-- $Id: books.xml 435 2008-01-20 02:15:28Z andre $ -->
<Books xmlns="">
  <Book isbn="2841771210">
    <Title>Developpement d'applications avec Objective Caml</Title>
      <Author>Emmanuel Chailloux</Author>
      <Author>Pascal Manoury</Author>
      <Author>Bruno Pagano</Author>
  <Book isbn="0201889544">
    <Title>C++ Programming Language</Title>
      <Author>Bjarne Stroustrup</Author>
  <Book isbn="0201710897">
    <Title>Programming Ruby</Title>
      <Author>David Thomas</Author>
      <Author>Andrew Hunt</Author>

books_error.xml (Non valid XML)

<?xml version="1.0" encoding="UTF-8"?>
<!-- $Id: books_error.xml 436 2008-01-20 02:33:53Z andre $ -->
<Books xmlns="">
  <Book isbn="2841771210">
    <Title>Developpement d'applications avec Objective Caml</Title>
      <Author>Emmanuel Chailloux</Author>
      <Author>Pascal Manoury</Author>
      <Author>Bruno Pagano</Author>
  <!-- Intentional error in the isbn format:
  3 digits instead of 10 (see XML Schema books.xsd) -->
  <Book isbn="123">
    <Title>C++ Programming Language</Title>
      <Author>Bjarne Stroustrup</Author>
  <Book isbn="0201710897">
    <Title>Programming Ruby</Title>
      <Author>David Thomas</Author>
      <Author>Andrew Hunt</Author>

win32com is a module from the Python for Windows Extensions, it wraps the Windows COM API's. The win32com.client.Dispatch() method creates COM objects from their ProgID's. The 2 ProgID's used are respectively: Msxml2.DOMDocument.6.0 (DOM document), and Msxml2.XMLSchemaCache.6.0 (schemas collection). You may have to use a different version if you have only MSXML 4.0 installed. After creating the DOM document COM object, set its async property to false (dom.async = 0) to use the load method in synchronous mode. Create a schema collection and add the XML schema and its namespace if any (if no namespace, an empty string is used instead). Assign the schema collection to the DOM document instance and load the document. The load method returns True if the validation is successful:

>>> import win32com.client
>>> dom = win32com.client.Dispatch('Msxml2.DOMDocument.6.0')
>>> dom.async = 0
>>> schemas = win32com.client.Dispatch('Msxml2.XMLSchemaCache.6.0')
>>> schemas.add('', 'books.xsd')
>>> dom.schemas = schemas
>>> dom.load('books.xml')

If the validation fails, the load method returns False. In addition the parseError property of the DOM document contains information about the error:

>>> dom.load('books_error.xml')
>>> dom.parseError.errorCode
>>> dom.parseError.reason
u"'020188954' violates pattern constraint of '[0-9]{10}'.\r\nThe attribute 'isbn' with value '020188954' failed to parse.\r\n"

Complete Python Example

"""Validate an XML file (1st param) against an XML schema
(2nd param), Using MSXML and PyWin32.

Copyright 2008 (c) Andre Burgaud
MIT License

Version 1.1

$Id: 435 2008-01-20 02:15:28Z andre $

from win32com.client import Dispatch
import os
import sys
from pywintypes import com_error

%s: Validation Error
- Error Code : %s
- Reason     : %s
- Character  : %s
- Line       : %s
- Column     : %s
- Source     : %s

Namespace : %s
Schema    : %s
Valid XML : %s

# MSXML versions to use
MSXML_DOM = 'Msxml2.DOMDocument.%d.0'
MSXML_SCHEMAS = 'Msxml2.XMLSchemaCache.%d.0'

def find(xml_file, xsd_file):
  """Validate existence of files."""
  for xfile in (xml_file, xsd_file):
    if not os.path.exists(xfile):
      print 'File %s not found...' % xfile
      return False
  return True

class MsXmlValidator:
  """Encapsulate some basic MXSML functionalities to validate an XML
  file against an XML Schema.

  def __init__(self):
    """Instantiate DOMDocument and XMLSchemaCache."""
    self.dom = self.create_dom()
    self.schemas = self.create_schemas()
    self.msxml_version = 0

  def create_dom(self):
    """Create and return a DOMDocument."""
    versions = MSXML_VERSIONS
    dom = None
    # Loop true the possible MSXML versions
    # Pick the first one created successfully
    for version in versions:
        dom = Dispatch(MSXML_DOM % version)
        self.msxml_version = version
        print 'MSXML version %s: OK' % version
      except com_error, msg:
        print msg
        print 'MSXML version %d: problem on this system' % version
    if not dom:
      print 'No compatible MSXML versions found on this system'
    dom.async = 0 # false
    return dom

  def create_schemas(self):
    """Create and return a Schema Cache."""
    schemas = Dispatch(MSXML_SCHEMAS % self.msxml_version)
    return schemas

  def get_namespace(self, xsd_file):
    """Extract targetNamespace, if any,  from XML schema."""
    namespace = ''
    self.dom.setProperty("SelectionLanguage", "XPath")
    path = "/*/@targetNamespace"
    node = self.dom.documentElement.selectSingleNode(path)
    if node:
      namespace = node.text
    return namespace

  def add_schema(self, namespace, xsd_file):
    """Add schema and namespace to Schema Cache."""
      self.schemas.add(namespace, xsd_file)
    except com_error, msg:
      print 'Error in XML Schema: %s' % xsd_file
      print msg
    self.dom.schemas = self.schemas

  def validate_xml_file(self, xml_file, xsd_file):
    """Validate XML file against XML Schema file."""
    if not find(xml_file, xsd_file):
    namespace = self.get_namespace(xsd_file)
    self.add_schema(namespace, xsd_file)
    if self.dom.load(xml_file):
      print FORMAT_SUCCESS % (namespace, xsd_file, xml_file)
      error = self.dom.parseError
      print FORMAT_ERROR % (xml_file,
                             " " * (error.linepos - 1) + '^')

def main():
  """Handle parameters, create Validator and invoke validation."""
  if len(sys.argv) < 3:
    print 'Usage: %s xml_file xsd_file' % sys.argv[0]
  validator = MsXmlValidator()
  xml_file, xsd_file = sys.argv[1], sys.argv[2]
  validator.validate_xml_file(xml_file, xsd_file)

if __name__ == '__main__':

The namespace is discovered from the XML schema file, using the XPath feature from MSXML. In a real context this approach might be questionable and the namespace could be provided as a parameter to the script for example.

This script takes 2 parameters (XML file and XML Schema file). Here is an example of execution with the valid XML file:

C:\prompt>python books.xml books.xsd
MSXML version 6: OK
Namespace :
Schema    : books.xsd
Valid XML : books.xml

And an example with the non valid XML:

C:\prompt>python books_error.xml books.xsd
MSXML version 6: OK
books_error.xml: Validation Error
- Error Code : -1072897687
- Reason     : '123' violates pattern constraint of '[0-9]{10}'.
The attribute 'isbn' with value '123' failed to parse.
- Character  : 534
- Line       : 14
- Column     : 20
- Source     :   <Book isbn="123">



The source code accompanying this post is released under the MIT License.

Release Notes

  • 01/26/2008: Updated source code and companion files (version 1.1)
  • 01/19/2008: Initial publication of this blog post.
Blog Comments powered by Disqus.

Next Post Previous Post