Re: Having trouble with siblings

Same tricks from Re: Get Node Value from irregular XML (xpather.pl) "//header[ contains(.,'rationale') ]/* "

xmllint.exe --xpath " //header[ contains(.,'rationale') ]/l /preceding
+-sibling::* " fudge
xmllint.exe --xpath " //header[ contains(.,'rationale') ]/child::text(
+) " fudge
[download]

where fudge is your data part=1 part=2

<fudge> 
    <root>
        <part>
            <sect>
               <header>
                   1. Purpose and rationale
                   <p>purpose 1</p>
                   <p>Purpose 2</p>
                   <p>purpose 3</p>
                   <l>
                      <li>purpose list 1</li>
                      <li>list2</li>
                    </l>                      
                </header>
            </sect>
        </part>
    </root>
    <root>
        <part>
            <sect>
                <header>
                    2. Purpose and rationale
                    <p>2 purpose 1</p>
                    <p>2 Purpose 2</p>
                    <p>2 purpose 3</p>
                    <l>
                        <li>2 purpose list 1</li>
                        <li>2 list2</li>
                    </l>                      
                    <p>2. Some other heading</p>
                    <p>2 content 1</p>
                    <p>2 content 2</p>
                </header>
            </sect>
        </part>
    </root>
</fudge>
[download]

The two queries combined, with the results

xmllint.exe --xpath " //header[ contains(.,'rationale') ]/child::text(
+) | //header[ contains(.,'rationale') ] /l/preceding-sibling::*   " f
+udge


                   1. Purpose and rationale
                   <p>purpose 1</p>
                   <p>Purpose 2</p>
                   <p>purpose 3</p>


                    2. Purpose and rationale
                    <p>2 purpose 1</p>
                    <p>2 Purpose 2</p>
                    <p>2 purpose 3</p>
[download]

Comment on Re: Having trouble with siblings Select or Download Code

Replies are listed 'Best First'.

Re^2: Having trouble with siblings
by Anonymous Monk on Jun 30, 2013 at 15:34 UTC

Aha, the node test node() will select all types of nodes, even text() nodes

so find a <header> which contains 'rationale'
and select every child node of that header
and filter this nodeset for nodes which have following-sibling with tagname <l> or which have tagname <l>

 xmllint.exe --xpath " //header[ contains(.,'rationale') ]/node()[ fol
+lowing-sibling::l or self::l]   " fudge

                   1. Purpose and rationale
                   <p>purpose 1</p>
                   <p>Purpose 2</p>
                   <p>purpose 3</p>
                   <l>
                      <li>purpose list 1</li>
                      <li>list2</li>
                    </l>
                    2. Purpose and rationale
                    <p>2 purpose 1</p>
                    <p>2 Purpose 2</p>
                    <p>2 purpose 3</p>
                    <l>
                        <li>2 purpose list 1</li>
                        <li>2 list2</li>
                    </l>
[download]

Naturally if there are more than one <l> you can limit to first <l> with position() xmllint.exe --xpath " //header[ contains(.,'rationale') ]/node()[ following-sibling::l[ position()=1 ] or self::l[ position()=1 ] ] " fudge

[reply]
[d/l]
[select]


Come for the quick hacks, stay for the epiphanies.
	PerlMonks