XmlDocument.LoadXml and namespaces

[David Thielen]

Hi;

Once I have loaded an xml file into a DOM, is there any way to then get any
namespace declarations that were in the xml file? Or are those just discarded
as the DOM is populated?

And if you can't get it from the DOM, any other way to get the namespace(s)?
I'm assuming doing a SAX read which returns each element including NameSpaces
and pull it form that (and build up the DOM at the same time)?

In my case I get a Stream so I can't read it twice as the Stream may not
support reset. And it may have NS elements - but I don't know if there are
any or what they are except by reading the Stream.

--
thanks - dave

[Oleg Tkachenko [MVP]]

David Thielen wrote:

Once I have loaded an xml file into a DOM, is there any way to then get any
namespace declarations that were in the xml file? Or are those just discarded
as the DOM is populated?

Of course not. You can select namespace declarations using
doc.SelectNodes("//namespace::*").

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[David Thielen]

Hi;

This sort-of worked. I have 3 namespaces declared in the header of the file.
The SelectNodes("//namespace::*") gave me 36 nodes. I did not check each one
but it appeared to be the 3 namespaces repeated over and over.

Any idea why I didn't get just 3?

--
thanks - dave
"Oleg Tkachenko [MVP]" wrote:

David Thielen wrote:

Once I have loaded an xml file into a DOM, is there any way to then get any
namespace declarations that were in the xml file? Or are those just discarded
as the DOM is populated?

Of course not. You can select namespace declarations using
doc.SelectNodes("//namespace::*").

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[David Thielen]

For the xml:
<root xmlns="http://www.test.org" xmlns:sns="http://www.test.org/sub"
xmlns:mns="http://www.test.org/mini">
<data>
<items>
<item id="1">
<sns:subItem sid="11">dave</sns:subItem>
<sns:subItem sid="12">thielen</sns:subItem>
</item>
<item id="2">
<sns:subItem sid="21">shirley</sns:subItem>
</item>
</items>
<mns:more>dave thielen</mns:more>
</data>
</root>

The nodes returned are:
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace
mns=http://www.test.org/mini
sns=http://www.test.org/sub
{default}=http://www.test.org
xml=http://www.w3.org/XML/1998/namespace

??? - thanks - dave

[Martin Honnen]

David Thielen wrote:

This sort-of worked. I have 3 namespaces declared in the header of the file.
The SelectNodes("//namespace::*") gave me 36 nodes. I did not check each one
but it appeared to be the 3 namespaces repeated over and over.

Any idea why I didn't get just 3?

Because in the XPath data model any element has in scope namespace nodes
and that way if you for instance declare namespaces on the root element
and do not declare/redeclare namespaces on any child elements then
besides the root element all child elements also have those namespaces
in scope as nodes of their own.

--

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

[Oleg Tkachenko [MVP]]

David Thielen wrote:

This sort-of worked. I have 3 namespaces declared in the header of the file.
The SelectNodes("//namespace::*") gave me 36 nodes. I did not check each one
but it appeared to be the 3 namespaces repeated over and over.

Any idea why I didn't get just 3?

Select unique ones (and omit default "xml" namespace):
//namespace::*[name() != 'xml'][not(../../namespace::*=.)]

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[David Thielen]

Oh right, that makes a lot of sense -- thanks - dave
"Martin Honnen" wrote:

David Thielen wrote:

This sort-of worked. I have 3 namespaces declared in the header of the file.
The SelectNodes("//namespace::*") gave me 36 nodes. I did not check each one
but it appeared to be the 3 namespaces repeated over and over.

Any idea why I didn't get just 3?

Because in the XPath data model any element has in scope namespace nodes
and that way if you for instance declare namespaces on the root element
and do not declare/redeclare namespaces on any child elements then
besides the root element all child elements also have those namespaces
in scope as nodes of their own.

--

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

[David Thielen]

Hi;

This worked great but I don't understand it totally. I get the [name() !=
'xml']

But what does [not(../../namespace::*=.)] mean? I'm totally lost on how that
works.

Also, when you do xpath[a][b] is that saying both a and b must be true? Can
you also do //namespace::*[name() != 'xml'] and [not(../../namespace::*=.)]
as the same thing? Or am I not understanding this?

I'm asking the above questions because I just want to understand xpath better.

--
thanks - dave
"Oleg Tkachenko [MVP]" wrote:

David Thielen wrote:

This sort-of worked. I have 3 namespaces declared in the header of the file.
The SelectNodes("//namespace::*") gave me 36 nodes. I did not check each one
but it appeared to be the 3 namespaces repeated over and over.

Any idea why I didn't get just 3?

Select unique ones (and omit default "xml" namespace):
//namespace::*[name() != 'xml'][not(../../namespace::*=.)]

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[Oleg Tkachenko [MVP]]

David Thielen wrote:

But what does [not(../../namespace::*=.)] mean? I'm totally lost on how that
works.
In XPath data model namespace nodes are propagated/copied down the tree
- each node in a subtree has namespace node for each namespace in context.
This predicate checks that parent of the element bearing the namespace
node doesn't have such namespace node. Effectively that allows to select
only namespace nodes on elements where namespace is declared.

Also, when you do xpath[a][b] is that saying both a and b must be true?
Yes.
Can
you also do //namespace::*[name() != 'xml'] and [not(../../namespace::*=.)]
as the same thing? Or am I not understanding this?

Yes, that's the same.

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[David Thielen]

Hi;

Ok, that makes sense. To make sure I understand this:
.../../namespace::* is getting the grandparent of the node on (../../) and
then getting any nodes in that element that start with namespace:: - why not
.../namespace::* - isn't that the parent?

And then the =. is saying equal to the node presently on . being the node on.

Is that correct?

--
thanks - dave
"Oleg Tkachenko [MVP]" wrote:

David Thielen wrote:

But what does [not(../../namespace::*=.)] mean? I'm totally lost on how that
works.

In XPath data model namespace nodes are propagated/copied down the tree
- each node in a subtree has namespace node for each namespace in context.
This predicate checks that parent of the element bearing the namespace
node doesn't have such namespace node. Effectively that allows to select
only namespace nodes on elements where namespace is declared.

Also, when you do xpath[a][b] is that saying both a and b must be true?

Yes.

Can
you also do //namespace::*[name() != 'xml'] and [not(../../namespace::*=.)]
as the same thing? Or am I not understanding this?

Yes, that's the same.

--
Oleg Tkachenko [XML MVP, MCAD]
http://www.xmllab.net
http://blog.tkachenko.com

[Martin Honnen]

David Thielen wrote:

Ok, that makes sense. To make sure I understand this:
../../namespace::* is getting the grandparent of the node on (../../) and
then getting any nodes in that element that start with namespace:: - why not
../namespace::* - isn't that the parent?

It helps looking into the XPath spec which here
<http://www.w3.org/TR/xpath#namespace-nodes>
says about the namespace nodes:
"Each element has an associated set of namespace nodes, one for each
distinct namespace prefix that is in scope for the element (including
the xml prefix, which is implicitly declared by the XML Namespaces
Recommendation [XML Names]) and one for the default namespace if one is
in scope for the element. The element is the parent of each of these
namespace nodes"
thus if you have a predicate on
//namespace::*
then
..
gives you the element the namespace node belongs to and
../..
gives the parent node of that element so that the predicate
not(../../namespace::*=.)
compares the namespace node on one element with the namespace nodes
attached to the parent element.


--

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

[David Thielen]

Hi;

I think I've read the spec at least 5 times. And I have 4 books on XML/XPath
that I have gone through. But I still trip on stuff - and figure it's best to
ask and ask until I know I understand a point well. I appreciate everyone
explaining the why behind this.

As soon as you said the .. is the node containing the namespace I realized
why it's ../.. instead of .. - should have figured that out.

One thing that hit me while going through this, it seems that xpath
basically walks the tree node by node to determine if that node matches the
xpath. If it does, it adds that node to the node(s) to be returned and walks
on to the next one. (Obviously in implementation it does not walk nodes it
knows can't match.)

So it makes sense that you can use . in the xpath to represent the node you
are presently evaluating. But I hadn't really comprehended that before as a
SQL select has no similiar concept.

Anyways, thanks for the lesson - I think I understand this now and this is
definitely useful info.

--
thanks - dave
"Martin Honnen" wrote:

David Thielen wrote:

Ok, that makes sense. To make sure I understand this:
../../namespace::* is getting the grandparent of the node on (../../) and
then getting any nodes in that element that start with namespace:: - why not
../namespace::* - isn't that the parent?

It helps looking into the XPath spec which here
<http://www.w3.org/TR/xpath#namespace-nodes>
says about the namespace nodes:
"Each element has an associated set of namespace nodes, one for each
distinct namespace prefix that is in scope for the element (including
the xml prefix, which is implicitly declared by the XML Namespaces
Recommendation [XML Names]) and one for the default namespace if one is
in scope for the element. The element is the parent of each of these
namespace nodes"
thus if you have a predicate on
//namespace::*
then
..
gives you the element the namespace node belongs to and
../..
gives the parent node of that element so that the predicate
not(../../namespace::*=.)
compares the namespace node on one element with the namespace nodes
attached to the parent element.


--

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