Reading fragments / .net 2.0

[Oliver Sturm]

I have an XML file that contains fragments, meaning there's no root
element. Node names are in my own "test" namespace. Looks like this:

<test:info date="...">Cont ent</test:info>
<test:info date="...">Othe r content</test:info>
....

I have some .net 1.1 code that reads the nodes from the file one by one:

FileStream stream = new FileStream(@"fi le.xml", FileMode.Open);
XmlNameTable nt = new NameTable( );
XmlNamespaceMan ager nsMgr = new XmlNamespaceMan ager(nt);
nsMgr.AddNamesp ace("test", "http://sturmnet.org/ns/test");
XmlParserContex t context = new XmlParserContex t(nt, nsMgr, null,
XmlSpace.None);

XmlTextReader reader = new XmlTextReader(s tream, XmlNodeType.Ele ment,
context);

XmlDocument doc = new XmlDocument( );
while (!reader.EOF) {
XmlNode node = doc.ReadNode(re ader);
// do something with the node
}

Now, using .net 2.0, docs say I'm supposed to have my XmlReader created
by XmlReader.Creat e instead of creating it myself. I was also hoping to
find a replacement for the reading of the nodes, because using that
XmlDocument, that's useless otherwise, just doesn't look like the most
elegant solution.

I fiddled around with this for a while, starting with

XmlReaderSettin gs settings = new XmlReaderSettin gs( );
settings.Confor manceLevel = ConformanceLeve l.Fragment;
XmlReader reader = XmlReader.Creat e(@"file.xml", settings);
while (!reader.EOF) {
reader.Read();
// now where's my node?
}

First problem: I get an error message saying the "test" namespace is
unknown and I have no idea where I'm supposed to stick my namespace in
this model.

So, what's the PC way to get this done in .net 2.0?
Oliver Sturm
--
omnibus ex nihilo ducendis sufficit unum
Spaces inserted to prevent google email destruction:
MSN oliver @ sturmnet.org Jabber sturm @ amessage.de
ICQ 27142619 http://www.sturmnet.org/blog

[Martin Honnen]

Oliver Sturm wrote:

I have an XML file that contains fragments, meaning there's no root
element. Node names are in my own "test" namespace. Looks like this:

<test:info date="...">Cont ent</test:info>
<test:info date="...">Othe r content</test:info>
...

I have some .net 1.1 code that reads the nodes from the file one by one:

FileStream stream = new FileStream(@"fi le.xml", FileMode.Open);
XmlNameTable nt = new NameTable( );
XmlNamespaceMan ager nsMgr = new XmlNamespaceMan ager(nt);
nsMgr.AddNamesp ace("test", "http://sturmnet.org/ns/test");
XmlParserContex t context = new XmlParserContex t(nt, nsMgr, null,
XmlSpace.None);

XmlTextReader reader = new XmlTextReader(s tream, XmlNodeType.Ele ment,
context);

XmlDocument doc = new XmlDocument( );
while (!reader.EOF) {
XmlNode node = doc.ReadNode(re ader);
// do something with the node
}

Now, using .net 2.0, docs say I'm supposed to have my XmlReader created
by XmlReader.Creat e instead of creating it myself. I was also hoping to
find a replacement for the reading of the nodes, because using that
XmlDocument, that's useless otherwise, just doesn't look like the most
elegant solution.

But if you want/need to have an XmlNode then you need to have an
XmlDocument as a node always belongs to its OwnerDocument.
--

Martin Honnen
http://JavaScript.FAQTs.com/

[Oliver Sturm]

Martin Honnen wrote:

But if you want/need to have an XmlNode then you need to have an
XmlDocument as a node always belongs to its OwnerDocument.

Yep. Well. Any specific reason why it would have to be that way?

Oliver Sturm
--
omnibus ex nihilo ducendis sufficit unum
Spaces inserted to prevent google email destruction:
MSN oliver @ sturmnet.org Jabber sturm @ amessage.de
ICQ 27142619 http://www.sturmnet.org/blog

[Oleg Tkachenko [MVP]]

Oliver Sturm wrote:

Now, using .net 2.0, docs say I'm supposed to have my XmlReader created
by XmlReader.Creat e instead of creating it myself. I was also hoping to
find a replacement for the reading of the nodes, because using that
XmlDocument, that's useless otherwise, just doesn't look like the most
elegant solution.

Your .NET 1.1 code still should work. I'm sure what about
XmlReader.Creat e() - looks like it doesn't cover proividing namespace
bindings when reading fragments.
And what for using XmlDocument - it's not clear which problem you are
trying to solve here.

--
Oleg Tkachenko [XML MVP, MCP]
http://blog.tkachenko.com

[Oliver Sturm]

Oleg Tkachenko [MVP] wrote:

Now, using .net 2.0, docs say I'm supposed to have my XmlReader
created by XmlReader.Creat e instead of creating it myself. I was also
hoping to find a replacement for the reading of the nodes, because
using that XmlDocument, that's useless otherwise, just doesn't look
like the most elegant solution.

Your .NET 1.1 code still should work.

Yes, it does. I'm just trying to comply to what the docs say. Of course
they also mention that there may be cases where an XmlTextReader still
needs to be constructed directly, so maybe this is one of them.
And what for using XmlDocument - it's not clear which problem you are
trying to solve here.

You mean, with using it, or with not wanting to use it? I'm using it the
way I am because I haven't found another way to read single nodes from a
file containing a sequence of fragments.

I don't really want to use it because it looks to me like a kludge,
creating an XmlDocument I don't really need just to use a single method
(that doesn't really have anything to do with the XmlDocument, AFAICS).
As Martin said in his post on this thread, currently there seems to be a
tight binding between the XmlNode and an XmlDocument, without any
apparent reason. I was thinking that maybe with all the restructuring
work that seems to have gone into System.Xml, maybe there'd be a more
elegant way to do this now.

Never mind, it's not a real problem. Thanks!
Oliver Sturm
--
omnibus ex nihilo ducendis sufficit unum
Spaces inserted to prevent google email destruction:
MSN oliver @ sturmnet.org Jabber sturm @ amessage.de
ICQ 27142619 http://www.sturmnet.org/blog