Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Indexing XML with a Custom XmlDocument

The XML Files — Thu, 12 Feb 2004 19:15:00 GMT

Indexer un document XML

Julien Cheyssial's Blog (FR) — Thu, 12 Feb 2004 21:02:00 GMT

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Dimitre Novatchev — Fri, 13 Feb 2004 12:37:00 GMT

This would be a very useful feature and will be best extended not only to store an id attribute's value but to generate and maintain an id for every type of node.

Yes, I need something with the functionality of the standard XSLT function generate-id() with time complexity much better than the O(N^2) exhibited by the current .Net XSLT implementation.

Therefore, why not just fix generate-id() itself and also the current xsl:key implementation?

Let me assure you that this is really very much needed by all XSLT programmers and not just an exercise for the curious of us.

Cheers,

Dimitre Novatchev [XML MVP],
FXSL developer, XML Insider,

http://fxsl.sourceforge.net/ -- the home of FXSL
Resume: http://fxsl.sf.net/DNovatchev/Resume/Res.html

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

RichB — Sat, 14 Feb 2004 13:23:00 GMT

{pedant}

You have used invalid URN syntax.

rfc2141 (URN Syntax) states:

All URNs have the following syntax (phrases enclosed in quotes are REQUIRED):

<URN> ::= "urn:" <NID> ":" <NSS>

{/pedant}

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Mark Fussell — Tue, 17 Feb 2004 06:34:00 GMT

Yes, the implementation of xsl:key in the .NET XSLT 1.0 and 1.1 processor had a bad bug. In the next Service Pack (SP) due to be released for .NET 1.0 and .NET 1.1 this has been fixed. I am sure that you know of this KB http://support.microsoft.com/default.aspx?scid=kb;EN-US;324478

The generate-id function although useful for generating id values for HTML anchors via href, cannot be used with the xsl:key or xsl:id functions which is a limitation. However adding a new user defined method to XSL to do this is a good idea. I will also look into the 0^2 behavior. As I mentioned at PDC 2003 we have done a lot of work to increase the perf of the XSLT processor in .NET V2.0

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Dimitre Novatchev — Wed, 18 Feb 2004 22:13:00 GMT

generate-id() *can* be used with xsl:key (e.g. in the "use" attribute) and correspondingly with the key() function and this use is not at all rare.

If you need concrete examples, I will be glad to provide them.

As for the coming fix, I knew this a year ago -- the problem is the fix is still not available.

Please, understand me correctly -- we are accustomed with the unprecedented quality and performance of MSXML4 and would expect the same from dotNet XslTransform, especially taking into account that it is much more easier to write an XSLT processor in a modern, garbage-collected language than in C++.

Thank you,

Dimitre Novatchev.

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Steve Saxon — Thu, 19 Feb 2004 04:52:00 GMT

This is a pretty neat way to work around the DTD/schema limitation.

Probably worth pointing out it would be really easy to build much the same thing for XPathDocument (now that the DOM is dead!).

You would derive from XPathDocument, and have a custom XmlReader that worked together with the reader. You hook the reader's Read() method to trap when you're on an element, and as one flys by, you look for the id attribute and pass it back to your XPathDocument-derived class through some internal method.

If you then added the XPath function extension to bind the id, that would work for this too, by the power of XPathNavigator.

BTW, the link for the XPath function extensions should be;
http://msdn.microsoft.com/library/en-us/dnexxml/html/xml10212002.asp

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

little things... — Thu, 19 Feb 2004 08:50:00 GMT

Missing from here:
- how to handle node removal from the document?
- how to handle node insertion in a document fragment?

re: Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class

Dimitre Novatchev — Thu, 19 Feb 2004 13:59:00 GMT

For a pure XSLT solution of a closely related problem -- to assign ids to nodes -- see the thread "Counting nodes efficiently" in the xsl-list:

http://aspn.activestate.com/ASPN/Mail/Message/XSL-List/2002866

The generated "_id" attribute can be easily used with a corresponding definition of an xsl:key and then with the key() function.

XML Bestiary: IndexingXPathNavigator

Signs on the Sand — Tue, 16 Mar 2004 16:21:00 GMT

Here I go again with another experimental creature from my XML Bestiary: IndexingXPathNavigator. This one was inspired by Mark Fussell's "Indexing XML, XML ids and a better GetElementByID method on the XmlDocument class". I've been experimenting with Mark's extended XmlDocument, played a bit with XPathDocument and "idkey()" extension function Mark...

xml:id gets First Working Draft

Signs on the Sand — Mon, 12 Apr 2004 14:49:00 GMT

In related news - The XML Core Working Group has released the First Public Working Draft of xml:id Version 1.0. The specification introduces a predefined attribute name that can always be treated as an ID and hence can always be recognized. What can be said? At last! Finally! xml:id Version...