XPath invalid expression when tag name has curly braces

Question

The name of the tag I am trying to get to is {http://whitehatsec.com/XML-api-Vuln}description. Conveniently every tag is prefixed with that lovely reference to the whitehat website. Unfortunately the xpath in lxml doesn't like it. I am currently trying vuln_root[0].xpath('//\{http://whitehatsec.com/XML-api-Vuln\}description') which should get me to the right node. However lxml keeps saying

File "test.py", line 23, in <module> for s in vuln_root[0].xpath('//\{http://whitehatsec.com/XML-api-Vuln\}description'): File "lxml.etree.pyx", line 1509, in lxml.etree._Element.xpath (src/lxml/lxml.etree.c:50725) File "xpath.pxi", line 318, in lxml.etree.XPathElementEvaluator.__call__ (src/lxml/lxml.etree.c:146020) File "xpath.pxi", line 238, in lxml.etree._XPathEvaluatorBase._handle_result (src/lxml/lxml.etree.c:145028) File "xpath.pxi", line 224, in lxml.etree._XPathEvaluatorBase._raise_eval_error (src/lxml/lxml.etree.c:144883) lxml.etree.XPathEvalError: Invalid expression

How can I get around this awful tag naming in my xpath? Thanks

`{namespace-uri}element-name` is one way to write an element name including its namespace (sometimes you'll also see these with a leading `Q` before the opening curly brace). It's the same as `ns-prefix:element-name` when you have a `xmlns:ns-prefix='namespace-uri'` declaration. — Charles Duffy, Jun 27 '14 at 22:01
For a bit more background: "It's a feature, not a bug!" -- XML namespaces let folks build compound documents without worrying about conflicts... so you could build a document that included both XHTML or XForms or whatever else and whitehat data, and not worry about whether any of the tag names conflict with each other. It's exceptionally useful combining things like XML template/macro languages with other document types, since the namespaces tell your editing tools for your larger documents that the markup content is foreign data and should be left alone. — Charles Duffy, Jun 27 '14 at 22:17

score 2 · Accepted Answer · answered Jun 27 '14 at 21:52

2

The easiest approach here is to just map the namespace.

el.xpath('//vuln:description',
    namespaces={'vuln': 'http://whitehatsec.com/XML-api-Vuln'})

answered Jun 27 '14 at 21:52

Charles Duffy

280,126
43
390
441

The vuln would need the {} around it. Is that still valid? – thaweatherman Jun 27 '14 at 21:53
1

No, it doesn't need the `{}` around it. The curly braces are part of the QName syntax. – Charles Duffy Jun 27 '14 at 21:54
Ok I see. I just was confused when the tag name included those. I'm not super familiar with xpath. It worked by the way. Initially typed namespaces without the last s. Thanks for your help! – thaweatherman Jun 27 '14 at 21:56
Not strictly an XPath thing, by the way -- you'll see the same notations in a number of other contexts involving XML namespaces as well. – Charles Duffy Jun 27 '14 at 22:02

XPath invalid expression when tag name has curly braces

1 Answers1