Hi Simon, Am Sonntag, 17. Juni 2012, 16:18:45 schrieb Simon Sapin:

I would like to get the namespace URI and local part of an element’s name separately. I think I could manage with the C/Cython API, but I don’t see how to do it from Python. The best I can think of is:

def split_ns(element): tag = element.tag if '}' in tag: ns, local = tag.split('}', 1) return ns[1:], local # remove '{' from ns else: return None, tag

Is the above correct? (Ie. can the URI or local part ever contain '}'?) Is there a better way?

I had the same problem, but solved it with a regular expression. If you are only interested in the local name, try this: ------------- import re qname=re.compile("{(?P<ns>.*)}(?P<element>.*)") def localname(node): """Returns the local name part of node""" m = qname.search(node.tag) return m.groupdict().get("element") if m else node.tag # We assume, there is a variable root which is an element # root = '{http://docbook.org/ns/docbook}book' print localname(root) # returns "book" ------------- The above localname() function works with and without a namespace. Either way, it returns the local name of the element name. If you need both, URI and its local name, use the "qname" variable--or modify the above function to return both parts. If there is a match, the qname.groupdict() returns a dictionary like this: {'ns': 'http://docbook.org/ns/docbook', 'element': 'book'} With a more clever regular expression, you can make the {...} part optional and still get a match: qname=re.compile("({(?P<ns>.*)})?(?P<element>.*)") m=qname.search("chapter") # just as an example m.groupdict() # There is no namespace, therefor 'ns' is empty! {'ns': None, 'element': 'chapter'} I don't think, there is a method to return the local name in the lxml API (at least, I haven't seen one). Hope that helps. -- Gruß/Regards Thomas Schraitle