XPath select node with periods

Question

I have an XML document where some of the nodes have a . in their name:

<com.site.blah>
   <id>asdkjasd</id>
   <com.site.testing>
       <name>test</name>
    </com.site.testing>
</com.site.blah>

If I try @doc.search("/*/id").first.xpath, it returns /com.site.blah/id, but if I then do: @doc.search("/com.site.blah/id").first.inspect it returns nil.

I want to be able to make an XPath query to select the name under com.site.testing, but it keeps rejecting my queries.

Any ideas?

(I am using hpricot if it makes a difference)

@Nikita - Yes, I have tried `com\.site\.blah` and that does not work — Mitch Dempsey, Aug 09 '10 at 20:37
Why didn't you try `com.site.blah`? (http://codepad.org/ecVVUPCu) — Tomalak, Aug 09 '10 at 20:38
@Tomalak - Nope, the XML document is basically exactly what I posted. There is no ` — Mitch Dempsey, Aug 09 '10 at 20:48
THis is actually *not* an XPath question but an xpathengines question. In standard XPath using names with dots is OK, if you have problems then your XPath engine is not compliant as per W3C spec. — Dimitre Novatchev, Aug 09 '10 at 20:57
@webdestroya: It should be no problem. In MSXSL, I can get "asdkjasd" with `/com.site.blah/id`. — , Aug 09 '10 at 20:58
Related: [Parse XML with JRuby (Hpricot?) with tags like ``](http://stackoverflow.com/questions/1396511) — Tomalak, Aug 09 '10 at 20:43

score 7 · Accepted Answer · answered Jul 04 '13 at 14:11

Your XPath library is broken. An XPath name test is a QName (http://www.w3.org/TR/xpath/#NT-NameTest), which, after following the rabbit hole of EBNF, includes periods (http://www.w3.org/TR/REC-xml/#NT-NameChar). Report the error to the developers.

As a workaround, neither of the escape mechanisms mentioned in the comments are part of XPath, but you can try using a predicate to check the element name, like so:

/*[name(.) = 'com.site.blah']/id

XPath select node with periods

1 Answers1

Linked