Can not get CDATA out of XML node using Nokogiri

rubyonrails-talk
1

Input appreciated: Trying to get data from a CDATA field out of an xml doc using Nokogiri.

This is the node: <ATTACHED_DOCUMENT> </ATTACHED_DOCUMENT>

When I query I am getting at best empty string, where I should expect the text “insert PDF code here”. This is what I have tried, without success:

xml_request_doc.xpath(“//attached_document”) returns: [#<Nokogiri::XML::Element:0x1252bfe name=“attached_document”>]

xml_request_doc.xpath(“//attached_document/text()”) returns:

xml_request_doc.xpath(“//attached_document”).value returns: NoMethodError Exception: undefined method `value’ for [#<Nokogiri::XML::Element:

xml_request_doc.xpath(“//attached_document”).inner_html returns: “”

Any thoughts where going wrong (I am successfully using Nokogiri for every other element in my doc.

Thx!

2

One way would be to check the children until you get the CData node:

parent_node = xml_request_doc.xpath("//attached_document")

# this assumes parent_node is not nil cdata_node = parent_node.children.detect {|n| n.cdata?}

# i think you can then just do: cdata_node.value

3

Thanks Gary, I had been trying the .children, and tried what you suggested to be sure, and I get nil as the return value for cdata_node = parent_node.children.detect {|n| n.cdata?}

It is very strange - the following is the xml in text before creating the Nokogiri doc, and it says <ATTACHED_DOCUMENT> has no children, and does not seem to be any other objects inside the attached_document node in the Nokogiri doc:

<USER_INFO login=\"aaa\" password=\"bbb\" /> <REQUEST request_type=\"NEW\"> <SUPPLEMENT_ORDER_DETAIL report_id=\"ccc\" ordered_by=\"ddd\" phone=\"eee\" email=\"fff\"       faxed_docs=\"ggg\" rush=\"T\" reason=\"hhh\">     <TRADE_SUPPLEMENT account_number=\"iii\" creditor=\"jjj\" included_bankruptcy=\"T\">       <ATTACHED_DOCUMENT> <![CDATA[zzz]]> </ATTACHED_DOCUMENT>    </TRADE_SUPPLEMENT> </SUPPLEMENT_ORDER_DETAIL> </REQUEST>"

4

maybe try .cdatas on the parent node rather than children. Not sure about Nokogiri but such a method exists for rexml/document.

5

maybe try .cdatas on the parent node rather than children. Not sure about Nokogiri but such a method exists for rexml/document.

Thanks. It does work with REXML, out of the box, no fiddling. I like Nokogiri's speed but am finding it harder to work with - maybe it is just getting used to and I don't understand things, but aside from this what else I find a little strange is when Nokogiri matches on a node/element it wants everything in lowercase, even if my xml is uppercase. Anyhow +1 for old REXML.

6

maybe try .cdatas on the parent node rather than children. Not sure about Nokogiri but such a method exists for rexml/document.

This is beautiful in REXML, just give me what I want, no negotiation -- back to the joy of Rails:

REXML::XPath.first(xml_request_doc, "//ATTACHED_DOCUMENT").cdatas