Question

Here’s a chunk of XML:

<xml>
  <section>
    <element>good</element>
    <section class="ignored">
      <element>bad</element>
    </section>
  </section>
</xml>

It’s easy enough to select all elements, or all elements inside a section.ignored:

@doc.css('element').text
 => "goodbad" 
@doc.css('section.ignored element').text
 => "bad" 

But how do I select all elements that are not inside section.ignored? This doesn’t work:

@doc.css('section:not(.ignored) element').text
 => "goodbad" 

…because that actually means “all elements that are contained in any section that is not ignored”, including the top-level section that wraps everything else!

Additional twist: unlike the simplified sample above, the real XML I have to deal with is nested to arbitrary depth, including sections within the ignored section(s).

And yes, I could just substract the bad array from the full array in Ruby and call it a day, but I’d prefer a pure CSS solution (or, if you must, XPath) if possible.

Answer 1

how do I select all elements that are not inside section.ignored ?

Use this XPath expression:

//element[not(ancestor::section[@class='ignored'])]

This selects any element named element that doesn’t have an ancestor named section, the string value of whose class attribute is the string "ignored" .

XSLT – based verification:

<xsl:stylesheet version="1.0"
 xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
 <xsl:output omit-xml-declaration="yes" indent="yes"/>
 <xsl:strip-space elements="*"/>

 <xsl:template match="/">
     <xsl:copy-of select=
     "//element[not(ancestor::section[@class='ignored'])]"/>
 </xsl:template>
</xsl:stylesheet>

When this transformation is applied on the provided XML document:

<xml>
    <section>
        <element>good</element>
        <section class="ignored">
            <element>bad</element>
        </section>
    </section>
</xml>

the above XPath expression is evaluated and all (in this case just one) selected nodes are copied to the output. The wanted, correct result is produced:

<element>good</element>