Possible using XPath?

Let's say I have the following structure:

<?xml version="1.0" encoding="UTF-8"?>
<root>
<child id="c1">
<child id="gc1">
<child id="ggc1"/>
<child id="ggc2"/>
</child>
<child id="gc2">
<child id="ggc3"/>
<child id="ggc4"/>
</child>
</child>
<child id="c2">
<child id="gc3">
<child id="ggc5"/>
<child id="ggc6"/>
</child>
<child id="gc4">
<child id="ggc7"/>
<child id="ggc8"/>
</child>
</child>
</root>

By using the following XPath query

//child[@id="gc3"]/child

I can get the child nodes of "gc1". But what I'd really like to get
is the sub branch/path going back to the root. So instead of just
returning the two nodes

<child id="ggc5"/>
<child id="ggc6"/>

I'd like to be able to return the sub branch/path

<root>
<child id="c2">
<child id="gc3">
<child id="ggc5"/>
<child id="ggc6"/>
</child>
</child>
</root>

Is that possible? Or is this something I'd have to do programatically
using the nodes returned by the XPath query? Basically, I'm just
trying to get a fragment of the larger xml document...

thnx,
Christoph

On 27/03/2008, Christoph Boget <christoph.boget@gmail.com> wrote:
> Let's say I have the following structure:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <root>
> <child id="c1">
> <child id="gc1">
> <child id="ggc1"/>
> <child id="ggc2"/>
> </child>
> <child id="gc2">
> <child id="ggc3"/>
> <child id="ggc4"/>
> </child>
> </child>
> <child id="c2">
> <child id="gc3">
> <child id="ggc5"/>
> <child id="ggc6"/>
> </child>
> <child id="gc4">
> <child id="ggc7"/>
> <child id="ggc8"/>
> </child>
> </child>
> </root>
>
> By using the following XPath query
>
> //child[@id="gc3"]/child
>
> I can get the child nodes of "gc1". But what I'd really like to get
> is the sub branch/path going back to the root. So instead of just
> returning the two nodes
>
> <child id="ggc5"/>
> <child id="ggc6"/>
>
> I'd like to be able to return the sub branch/path
>
> <root>
> <child id="c2">
> <child id="gc3">
> <child id="ggc5"/>
> <child id="ggc6"/>
> </child>
> </child>
> </root>
>
> Is that possible? Or is this something I'd have to do programatically
> using the nodes returned by the XPath query? Basically, I'm just
> trying to get a fragment of the larger xml document...


//child[@id='gc3']/child/ancestor-or-self::*


-robin

> > Is that possible? Or is this something I'd have to do programatically
> > using the nodes returned by the XPath query? Basically, I'm just
> > trying to get a fragment of the larger xml document...
> //child[@id='gc3']/child/ancestor-or-self::*

Thanks for the response. However, I must be doing something wrong
here. The test script below isn't doing what I'm expecting:

$xml = '<?xml version="1.0" encoding="UTF-8"?><root><child
id="c1"><child id="gc1"><child id="ggc1">Great Grand Child
1</child><child id="ggc2">Great Grand Child 2</child></child><child
id="gc2"><child id="ggc3">Great Grand Child 3</child><child
id="ggc4">Great Grand Child 4</child></child></child><child
id="c2"><child id="gc3"><child id="ggc5">Great Grand Child
5</child><child id="ggc6">Great Grand Child 6</child></child><child
id="gc4"><child id="ggc7">Great Grand Child 7</child><child
id="ggc8">Great Grand Child 8</child></child></child></root>';

$doc = new DOMDocument('1.0', 'UTF-8');
$doc->loadXML( $xml );

$xpath = new DOMXPath($doc);
$nodeList = $xpath->query("//child[@id='gc3']/child/ancestor-or-self::*");

echo 'Got list list of [' . $nodeList->length . '] nodes:<br>';
for ($i = 0; $i < $nodeList->length; $i++) {
echo $nodeList->item($i)->nodeValue . "<br>\n";
}

When I run the XPath query through XMLSpy, the correct nodes are
returning. However, when I run it through DOMXPath->query() (or
evaluate() for that matter), it doesn't seem to be returning the
correct nodes.

What's going wrong?

thnx,
Chris

On 27/03/2008, Christoph Boget <jcboget@yahoo.com> wrote:
> > > Is that possible? Or is this something I'd have to do programatically
> > > using the nodes returned by the XPath query? Basically, I'm just
> > > trying to get a fragment of the larger xml document...
> > //child[@id='gc3']/child/ancestor-or-self::*
>
>
> Thanks for the response. However, I must be doing something wrong
> here. The test script below isn't doing what I'm expecting:
>
> $xml = '<?xml version="1.0" encoding="UTF-8"?><root><child
> id="c1"><child id="gc1"><child id="ggc1">Great Grand Child
> 1</child><child id="ggc2">Great Grand Child 2</child></child><child
> id="gc2"><child id="ggc3">Great Grand Child 3</child><child
> id="ggc4">Great Grand Child 4</child></child></child><child
> id="c2"><child id="gc3"><child id="ggc5">Great Grand Child
> 5</child><child id="ggc6">Great Grand Child 6</child></child><child
> id="gc4"><child id="ggc7">Great Grand Child 7</child><child
> id="ggc8">Great Grand Child 8</child></child></child></root>';
>
> $doc = new DOMDocument('1.0', 'UTF-8');
> $doc->loadXML( $xml );
>
> $xpath = new DOMXPath($doc);
> $nodeList = $xpath->query("//child[@id='gc3']/child/ancestor-or-self::*");
>
> echo 'Got list list of [' . $nodeList->length . '] nodes:<br>';
> for ($i = 0; $i < $nodeList->length; $i++) {
> echo $nodeList->item($i)->nodeValue . "<br>\n";
> }
>

Only the nodes specified are in the list, but the *values* of the
those nodes include children that aren't in the list.

change your for-loop to this and you'll see just the expected nodes:

foreach ($nodeList as $node) {
echo $node->tagName, ' : ', $node->getAttribute('id'), "<br>\n";
}

does that make sense?

-robin