From n xmlReader to single XPathDocument?

[Keith Chadwick]

My problem is this. I have to create several xmlReader objects each
retrieving 'for xml' formatted sql server data. I then need to peice them
together into a single document and place them into a single XPathDocument
which is then transformed throughout the web site with different
XslTransformati ons. I realize that the Xpathdocument will accept a
XmlReader as an object but I have 5 of them so how to 'best' combine. The
resulting XpathDocument is placed into the Cache and is only updated twice a
day.

Cheers
Keith

Note: I do note have access to the SQLXML objects on the production ISP site
otherwise I would use them!

[Oleg Tkachenko]

Keith Chadwick wrote:

My problem is this. I have to create several xmlReader objects each
retrieving 'for xml' formatted sql server data. I then need to peice them
together into a single document and place them into a single XPathDocument
which is then transformed throughout the web site with different
XslTransformati ons. I realize that the Xpathdocument will accept a
XmlReader as an object but I have 5 of them so how to 'best' combine.

You can create custom XmlReader, which reads all 5 other XmlReaders. See
"Customized XML Reader Creation" chapter in .NET Framework Developer's
Guide at
http://msdn.microsoft.com/library/de...ercreation.asp.

A more generic approach is XInclude. You can have one master XML, e.g.
<document xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="chapter1. xml" />
<xi:include href="chapter2. xml" />
<xi:include href="chapter3. xml" />
</document>
and then just read this document to XPathDocument via XIncludingReade r
(I'm talking about XInclude.NET [1]).

And what for SQL logic - you can encapsulate it to a custom XmlResolver
and provide it to XIncludingReade r, then having master document like

<document xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
</document>

[1] http://workspaces.gotdotnet.com/xinclude
--
Oleg Tkachenko
XML Insider
http://www.tkachenko.com/blog

[Keith Chadwick]

Thanks Oleg,

I will read through and give it a try. The first example at MSDN is not
really worth it for what I have, seems like an awful lot of work for
something that should be relatively simple!

I am looking into the XInclude which I have downloaded.

Worst case scenario is I can load the sql/xml data and write it to file then
tell the XPathDocument to simply load it. In MSXML4.0 it was real easy,
guess I was expecting it to be as easy again. I am still struggling with
the objects under xsl/xml etc and how they work together. I find the
documentation somewhat lacking to say the least and the books I have only
cover very basic scenarios, of course!

By the way your suggestion to use the XPathDocument for my transformations
has greatly improved performance. I just ran an application test with 50
000 hits and had no problems what so ever, stratvucha, have no idea how to
spell thank you in russion only how to say it :-)

Cheers
Keith

"Oleg Tkachenko" <oleg@NO!SPAM!P LEASEtkachenko. com> wrote in message
news:Of******** *****@TK2MSFTNG P11.phx.gbl...

Keith Chadwick wrote:

My problem is this. I have to create several xmlReader objects each
retrieving 'for xml' formatted sql server data. I then need to peice them together into a single document and place them into a single XPathDocument which is then transformed throughout the web site with different
XslTransformati ons. I realize that the Xpathdocument will accept a
XmlReader as an object but I have 5 of them so how to 'best' combine.
You can create custom XmlReader, which reads all 5 other XmlReaders. See
"Customized XML Reader Creation" chapter in .NET Framework Developer's
Guide at

http://msdn.microsoft.com/library/de...ercreation.asp.
A more generic approach is XInclude. You can have one master XML, e.g.
<document xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="chapter1. xml" />
<xi:include href="chapter2. xml" />
<xi:include href="chapter3. xml" />
</document>
and then just read this document to XPathDocument via XIncludingReade r
(I'm talking about XInclude.NET [1]).

And what for SQL logic - you can encapsulate it to a custom XmlResolver
and provide it to XIncludingReade r, then having master document like

<document xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
</document>

[1] http://workspaces.gotdotnet.com/xinclude
--
Oleg Tkachenko
XML Insider
http://www.tkachenko.com/blog

[Keith Chadwick]

> And what for SQL logic - you can encapsulate it to a custom XmlResolver

and provide it to XIncludingReade r, then having master document like

<document xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
<xi:include href="sql://yoursqldata" />
</document>

Ok Oleg you wanna explain the above a little. I have the XInclude now but
could you provide an example of the magic that occurs with
"sql://yoursqldata". I hate black boxes and honestly do NOT understand the
hidden magic here.

Humbly
Keith

[Oleg Tkachenko]

Keith Chadwick wrote:

By the way your suggestion to use the XPathDocument for my transformations
has greatly improved performance. I just ran an application test with 50
000 hits and had no problems what so ever, stratvucha, have no idea how to
spell thank you in russion only how to say it :-)

Credits due to XPathDocument developers at Microsoft, who created such
greatly effective and innovative component.
--
Oleg Tkachenko
XML Insider
http://www.tkachenko.com/blog

[Oleg Tkachenko]

Keith Chadwick wrote:

Ok Oleg you wanna explain the above a little. I have the XInclude now but
could you provide an example of the magic that occurs with
"sql://yoursqldata". I hate black boxes and honestly do NOT understand the
hidden magic here.

Well, sorry Keith, it was only an idea. This is not implemented in the
XInclude.NET yet, but sure will be in the next release.

The idea is based on custom XmlResolver technique, widely used in Java
world (it's called URIResolver in JAXP). First you implement custom
XmlResolver, which understands some proprietary URI schema, say sql://,
register it in XIncludingReade r and use such proprietary URIs in your
master XML document, say
<xi:include href="sql://blah.blah.some. your.data"/>
XIncludingReade r will call your XmlResolver to resolve the URI in href
attribute.
So in the custom XmlResolver code whenever you've been asked to resolve
such URI, you make you sql/xml logic and return result as XmlReader object.
That's it. I've heard one very smart guy is preparing an article about
using custom XmlResolvers, so watch MSDN news.

Meanwhile you can try merging via temporary files.
--
Oleg Tkachenko
XML Insider
http://www.tkachenko.com/blog

[Oleg Tkachenko]

Oleg Tkachenko wrote:

Meanwhile you can try merging via temporary files.

Alternatively create some aspx page, which returns generated XML and
call it in master document:
<doc xmlns:xi="http://www.w3.org/2003/XInclude">
<xi:include href="http://www.contoso.com/genxml.aspx?nam e=foobar"/>
</doc>
--
Oleg Tkachenko
XML Insider
http://www.tkachenko.com/blog

[Stephen Cook]

Keith,

Far be it for me to provide an alternative to the industrious Oleg,
[Excellent contributions both here and on his weblog]

If you're not too concerned about performance and are trying to build a
hierarchical XML representation of some obvious SQL relationships it may be
easier to build 5 data adaptors, create the relationships and then extract
the XML.

By test code looks like...

Database connection string in strConn
Streamwriter in strmOutput
SqlConnection conn = new SqlConnection(s trConn);
SqlDataAdapter dtaHolder = new SqlDataAdapter( "Select * from LPS_HOLDER
where FIRST_NAME like '%cook%'",conn) ;
SqlDataAdapter dtaLocation = new SqlDataAdapter( "Select * from
LPS_LOCATION",c onn);
conn.Open();
DataSet dsHoldLoc = new DataSet("Holder Location");
dtaHolder.Fill( dsHoldLoc,"Hold ers");
dtaLocation.Fil l(dsHoldLoc, "Locations" );
conn.Close();

DataRelation drHoldLoc = dsHoldLoc.Relat ions.Add("HoldL ocs",
dsHoldLoc.Table s["Locations"].Columns["INITIAL_LOCATI ON_ID"],
dsHoldLoc.Table s["Holders"].Columns["INITIAL_LOCATI ON_ID"],false);
drHoldLoc.Neste d = true;

dsHoldLoc.Write Xml(strmOutput) ;
strmOutput.Flus h();
strmOutput.Clos e();

Stephen

"Keith Chadwick" <kc*******@leew ardsystems.com> wrote in message
news:ei******** ******@TK2MSFTN GP09.phx.gbl...

My problem is this. I have to create several xmlReader objects each
retrieving 'for xml' formatted sql server data. I then need to peice them
together into a single document and place them into a single XPathDocument
which is then transformed throughout the web site with different
XslTransformati ons. I realize that the Xpathdocument will accept a
XmlReader as an object but I have 5 of them so how to 'best' combine. The
resulting XpathDocument is placed into the Cache and is only updated twice a day.

Cheers
Keith

Note: I do note have access to the SQLXML objects on the production ISP site otherwise I would use them!

[Keith Chadwick]

This is basicaly what I did, but with datasets. 5 datasets merged into one
then pulled out as string which is then placed in an xmldocument which is in
turn placed in a xmlNodeReader which is then placed into the xpathdocument.
Works very well and although it might not be the most efficient means of
doing it, the resulting xpathdocument is placed in the cache and expires
each day at midnight. so efficinecy was is not my primary concern The
process only occurs once a day.

Cheers
Keith

"Stephen Cook" <st************ ****@agilisysNO SPAM.co.uk> wrote in message
news:u5******** ******@TK2MSFTN GP11.phx.gbl...

Keith,

Far be it for me to provide an alternative to the industrious Oleg,
[Excellent contributions both here and on his weblog]

If you're not too concerned about performance and are trying to build a
hierarchical XML representation of some obvious SQL relationships it may be easier to build 5 data adaptors, create the relationships and then extract
the XML.

By test code looks like...

Database connection string in strConn
Streamwriter in strmOutput
SqlConnection conn = new SqlConnection(s trConn);
SqlDataAdapter dtaHolder = new SqlDataAdapter( "Select * from LPS_HOLDER
where FIRST_NAME like '%cook%'",conn) ;
SqlDataAdapter dtaLocation = new SqlDataAdapter( "Select * from
LPS_LOCATION",c onn);
conn.Open();
DataSet dsHoldLoc = new DataSet("Holder Location");
dtaHolder.Fill( dsHoldLoc,"Hold ers");
dtaLocation.Fil l(dsHoldLoc, "Locations" );
conn.Close();

DataRelation drHoldLoc = dsHoldLoc.Relat ions.Add("HoldL ocs",
dsHoldLoc.Table s["Locations"].Columns["INITIAL_LOCATI ON_ID"],
dsHoldLoc.Table s["Holders"].Columns["INITIAL_LOCATI ON_ID"],false);
drHoldLoc.Neste d = true;

dsHoldLoc.Write Xml(strmOutput) ;
strmOutput.Flus h();
strmOutput.Clos e();

Stephen

"Keith Chadwick" <kc*******@leew ardsystems.com> wrote in message
news:ei******** ******@TK2MSFTN GP09.phx.gbl...

My problem is this. I have to create several xmlReader objects each
retrieving 'for xml' formatted sql server data. I then need to peice them together into a single document and place them into a single XPathDocument which is then transformed throughout the web site with different
XslTransformati ons. I realize that the Xpathdocument will accept a
XmlReader as an object but I have 5 of them so how to 'best' combine. The resulting XpathDocument is placed into the Cache and is only updated

twice a

day.

Cheers
Keith

Note: I do note have access to the SQLXML objects on the production ISP

site

otherwise I would use them!