Saxon » SaxonC

Yes this seems like an oversight because SaxonC does not have a QName class like SaxonJ so there is no way to get the namespace prefix.

I have created a bug issue for this #6591 so hopefully we can get a fix out for this in the next maintenance release.

With the help of XPath prefix-from-QName(node-name($your-node)) you can get the prefix as a workaround as long as there is no QName equivalent in the SaxonC API, for instance with Python

from saxonche import *

xml1 = '''<root xmlns="http://example.com">
  <pf1:foo xmlns:pf1="http://example.com/ns2">test</pf1:foo>
  <pf2:bar xmlns:pf2="http://example.com/ns3">test2</pf2:bar>
</root>'''

def get_prefix(xdm_node : PyXdmNode, xpath_processor : PyXPathProcessor):
    xpath_processor.set_context(xdm_item=xdm_node)
    return xpath_processor.evaluate_single('prefix-from-QName(node-name())')

with PySaxonProcessor() as saxon_processor:
    print(saxon_processor.version)

    xdm_node = saxon_processor.parse_xml(xml_text=xml1)

    xpath_processor = saxon_processor.new_xpath_processor()

    xpath_processor.set_context(xdm_item=xdm_node)

    elements = xpath_processor.evaluate('//*')

    for element in elements:
        print(element.name)
        print(f'prefix: {get_prefix(element, xpath_processor)}')

Thanks for the workaround Martin. Unfortunately we don't have python on our target system; so we'll have to wait for a fix.

An equivalent workaround can be done using the C++ API. Same logic really using XPathProcessor class

Oh. Ok. Thanks O'Neil. I'll give it a try.

I was able to put together a simple C++ test program (based on Martin's python example) that gets the namespace prefixes for all the elements in a document. Thanks for your help on this guys.