Why can't I get a result from an XPath with namespace in the root element? [duplicate] - ruby

Question

This question already has answers here:
Nokogiri/Xpath namespace query
(3 answers)
Closed 8 years ago.
This is probably an XML namespace newbie question but I can't figure out how to get an XPath to work with the following trunctated XML with this particular root element:
<?xml version="1.0" encoding="UTF-8"?>
<CreateOrUpdateEventsRequest xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://dhamma.org" version="3-0-0">
<LanguageKey>
<IsoCode>en</IsoCode>
</LanguageKey>
<Publish>
<Value>true</Value>
</Publish>
<Events>
<Event>
<EventKey>
<LocationKey>
<SubDomain>rasmi</SubDomain>
</LocationKey>
<EventId>10DayPDFStdTag</EventId>
</EventKey>
</Event>
</Events>
</LanguageKey>
</CreateOrUpdateEventsRequest>
Using Ruby and Nokogiri (with a just updated libxml2), it works fine with XPath only if I delete all the extra info in the root element, making it:
<CreateOrUpdateEventsRequest>
Otherwise nothing works:
$> #doc.xpath("//CreateOrUpdateEventsRequest") #=> [] with original header, an array of nodes with modified header
$> #doc.xpath("//LanguageKey") #=> [] with the original header, an array of nodes with modified header
$> #doc.xpath("//xmlns:LanguageKey") #=> undefined namespace prefix with the original
How do I address namespaces like this with XPath?

Answer 1

Many thanks for the help.
The answer seems to be that the XML re-declared XMLNS when it should have declared the namespace with a prefix as in xmlns:myns.
From www.w3.org:
The XML specification reserves all names beginning with the letters 'x', 'm', 'l' in any combination of upper- and lower-case for use by the W3C. To date three such names have been given definitions—although these names are not in the XML namespace, they are listed here as a convenience to readers and users:
xml: See http://www.w3.org/TR/xml/#NT-XMLDecl and http://www.w3.org/TR/xml-names/#xmlReserved
xmlns: See http://www.w3.org/TR/xml-names/#ns-decl
xml-stylesheet: See The xml-stylesheet processing instruction

Answer 2

I don't use Nokogiri nor Ruby,
but you need to register a prefix for namespace http://dhamma.org
When I read http://nokogiri.org/tutorials/searching_a_xml_html_document.html
I understand you must do something like
$> #doc.xpath('//dha:LanguageKey', 'dha' => 'http://dhamma.org')

Answer 3

Here's some code to consider. Starting with code to create a Nokogiri::XML::Document:
require 'nokogiri'
XML = <<EOT
<?xml version="1.0" encoding="UTF-8"?>
<CreateOrUpdateEventsRequest xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://dhamma.org" version="3-0-0">
<LanguageKey>
<IsoCode>en</IsoCode>
</LanguageKey>
<Publish>
<Value>true</Value>
</Publish>
<Events>
<Event>
<EventKey>
<LocationKey>
<SubDomain>rasmi</SubDomain>
</LocationKey>
<EventId>10DayPDFStdTag</EventId>
</EventKey>
</Event>
</Events>
</LanguageKey>
</CreateOrUpdateEventsRequest>
EOT
doc = Nokogiri::XML(XML)
Here's the root node's name:
doc.root.name # => "CreateOrUpdateEventsRequest"
The docs say:
When using CSS, if the namespace is called “xmlns”, you can even omit the namespace name.
doc.at('CreateOrUpdateEventsRequest').name # => "CreateOrUpdateEventsRequest"
doc.at('LanguageKey').to_xml # => "<LanguageKey>\n <IsoCode>en</IsoCode>\n </LanguageKey>"
Using XPath, we can specify the default namespace as:
doc.at('//xmlns:LanguageKey').to_xml # => "<LanguageKey>\n <IsoCode>en</IsoCode>\n </LanguageKey>"
Sometimes, if there are a lot of namespaces it makes sense to use collect_namespaces and pass them in:
name_spaces = doc.collect_namespaces # =>
doc.at('//xmlns:LanguageKey', name_spaces).to_xml # => "<LanguageKey>\n <IsoCode>en</IsoCode>\n </LanguageKey>"
You'll need to look through the documentation for Nokogiri::XML::Node for more information on the various methods.
I recommend using CSS selectors for simplicity and readability over XPath, as a first try. I think XPath has more functionality but it makes my eyes bug out sometimes, so I prefer CSS.