How to tell if XmlNode is part of document tree

[martinfrompi]

I have a container of XmlNodes. Some of them have been removed from
the document, some haven't. Is there an easy way to tell which are
which?

I suppose
bool IsPartOfTree(Xm lNode node)
{
while (node.ParentNod e != null) { node = ParentNode; }
return node == node.OwnerDocum ent;
}
would work, but it doesn't seem elegant.

[Martin Honnen]

ma**********@ya hoo.co.uk wrote:

bool IsPartOfTree(Xm lNode node)
{
while (node.ParentNod e != null) { node = ParentNode; }

Should be
while (node.ParentNod e != null) { node = node.ParentNode ; }
I guess.

return node == node.OwnerDocum ent;
}
would work, but it doesn't seem elegant.

I don't think there is any method exposed that you could use so that
walk of the ParentNode hierarchy is not too bad.

If you don't want to use the loop you can use XPath e.g.

public static bool IsPartOfDocumen tTree (XmlNode node) {
return node.SelectSing leNode(@"ancest or::node()") ==
node.OwnerDocum ent;
}

--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[martinfrompi]

Martin Honnen wrote:

ma**********@ya hoo.co.uk wrote:

bool IsPartOfTree(Xm lNode node)
{
while (node.ParentNod e != null) { node = node.ParentNode ; }

[snip correction - thanks!]

return node == node.OwnerDocum ent;
}
would work, but it doesn't seem elegant.

I don't think there is any method exposed that you could use so that
walk of the ParentNode hierarchy is not too bad.

Thanks. Sigh. (I suppose there isn't much that an API could do, other
than duplicate the loop).

If you don't want to use the loop you can use XPath e.g.

public static bool IsPartOfDocumen tTree (XmlNode node) {
return node.SelectSing leNode(@"ancest or::node()") ==
node.OwnerDocum ent;
}

Hmm. That seems like a sledgehammer to crack a nut. It also feels
hairy applying XPath to something that isn't (part of) a full document.

(On a different topic - I really dislike the way that
SelectNodes/SelectSingleNod e ignores the axis ordering. I think I
would make the XPath expression:
"ancestor::node ()[last()]"
which ought to accurately select the top of the tree.)

[Keith Patrick]

Why are you walking up the parent tree to look for the OwnerDocument? All
descendant nodes, not just the root, reference the parent document. For
instance, the following (assuming my source document had enough children to
return nodes from the XPath) would output the owning document's ToString for
each node:
foreach (XmlNode node in
myDocument.Sele ctNodes("//*[1]/*[1]/*"))
{
System.Diagnost ics.Debug.Write Line(node.Owner Document);
}

[Martin Honnen]

Keith Patrick wrote:

Why are you walking up the parent tree to look for the OwnerDocument?

Because he needs to find out whether walking up the parent hierarchy he
finds the OwnerDocument as an ancestor or not to check whether the node
he started with is included in the tree rooted by the OwnerDocument or
in an orphan subtree detached from the OwnerDocument.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[martinfrompi]

On Jan 23, 2:18 pm, martinfro...@ya hoo.co.uk wrote:

I have a container of XmlNodes. Some of them have been removed from
the document, some haven't. Is there an easy way to tell which are
which?

I suppose
bool IsPartOfTree(Xm lNode node)
{
while (node.ParentNod e != null) { node = ParentNode; }
return node == node.OwnerDocum ent;
}
would work, but it doesn't seem elegant.

Oops. There are in fact TWO problems with the above code. Firstly, as
Martin Nonnen points out below, it should be:
node = node.ParentNode ;
secondly, the OwnerDocument property of the document node returns null.
So the final test should be:
return (node.NodeType == XmlNodeType.Doc ument);

So, in case somebody wants the full code, it should be:
public static bool IsPartOfDocumen tTree (XmlNode node)
{
while (node.ParentNod e != null)
{
node = node.ParentNode ;
}
return (node.NodeType == XmlNodeType.Doc ument);
}

(Note: The above has still not been compiled, let alone tested!)

[Keith Patrick]

I've never had a detached subtree still retain its reference to the original
document, so that's what I'm not seeing. Is he saying that when RemoveNode
is called, the doc reference remains?

[Dimitre Novatchev]

Using the Kaysian method for intersection of node-sets the intersection is
non-empty if and only if:
count($ns1[count(.|$ns2) = count($ns2)]) 0

We can use this to tell whether two nodes belong to the same document, by
defining the two node-sets as the sets of ancestors of each of the given
nodes:

Therefore, two nodes belong to the same document if and only if:
count($node1/ancestor::node( )[count(.|$node2/ancestor::node( )) =
count($node2/ancestor::node( )) ]) 0

To answer your specific question, represent the document by the document
node (or by the top element) in the above XPath expression.

This was tested successfully with the XPath Visualizer and it is confirmed
that the above expression returns 'true' only in the case when the two nodes
belong to the same XML document

Cheers,
Dimitre Novatchev

<ma**********@y ahoo.co.ukwrote in message
news:11******** **************@ q2g2000cwa.goog legroups.com...

>I have a container of XmlNodes. Some of them have been removed from
the document, some haven't. Is there an easy way to tell which are
which?

I suppose
bool IsPartOfTree(Xm lNode node)
{
while (node.ParentNod e != null) { node = ParentNode; }
return node == node.OwnerDocum ent;
}
would work, but it doesn't seem elegant.