Here is a Unicode question in foo/bär terms? You can change the text of an element from 'bär' to 'bar'. Why can't you change it from 'bar' to 'bär'? Or what does it take to do it? Here is a minimal file demonstrating the problem # This Python file uses the following encoding: utf-8 from lxml import etree import re text = """<root> <w>foo</w> <w>bär</w> </root>""" tree = etree.XML(text) for element in tree.iter('w'): if element.text.encode('utf-8') =='bär': print etree.tostring(element, encoding = 'utf-8') In the above first iteration of the script, the text 'bär' properly passes through the 'fromstring' and 'tostring' routines. In the next iteration of the script, you want to change 'bär' to 'bar'. That works for element in tree.iter('w'): if element.text.encode('utf-8') =='bär': element.text = 'bar' print etree.tostring(element, encoding = 'utf-8') That script returns the expected '<w>bar</w> But if I start with '<w>bar</w>' and want to change it to '<w>bär</w>', the script doesn't work: for element in tree.iter('w'): if element.text.encode('utf-8') =='bar': element.text = 'bär' print etree.tostring(element, encoding = 'utf-8') The script triggers the following error message: "ValueError: All strings must be XML compatible: Unicode or ASCII, no NULL bytes or control characters" I've tried various "recognize Unicode" routines that work in other contexts. But they don't work here. What am I missing? Martin Mueller Professor of English and Classics Northwestern University