answer for stackoverflow

q50550830.py

import lxml.etree as ET
import csv

# load file
tree = ET.parse('users.xml')
# iterate through each user tag
users = tree.findall('.//user')
all_roles = list({role.find('name').text for role in tree.findall('.//role')})
# just w mode, no wb. wb for binary data
with open('user_list.csv', "w") as csv_file:
    writer = csv.writer(csv_file, delimiter=',')
    # write headers
    writer.writerow([
        'email', 'enabled', 'externallyDefined',
        'fullName', 'tenantId', 'username', *all_roles
        ])
    for user in users:
        email = user.find('emailAddress').text
        enabled = user.find('enabled').text
        
        # process optional element
        externallyDefined = user.find('externallyDefined')
        if externallyDefined is not None:
            externallyDefined = externallyDefined.text

        fullName = user.find('fullName').text
        
        # another optional element
        tenantId = user.find('tenantId')
        if tenantId is not None:
            tenantId = tenantId.text

        username = user.find('username').text
        
        # collect nested elements (roles)
        user_roles = {}
        roles = user.find('roles').findall('role')
        for role in roles:
            user_roles[role.find('name').text] = True

        writer.writerow([
            email, enabled, externallyDefined, fullName,
            tenantId,  username, *[user_roles.get(role) for role in all_roles]
            ])