Hi,

On Don, 28 Sep 2000, you wrote:
> Why a DTD and not a Schema?  Just curious.

It was our first try - today I would use Schema to define the tag set.

> >   the one thing i would like to see is an
> > <xupdate:update-or-insert select="..">
> >   alternatively we can also include an
> > <xupdate:choose><xupdate:when ...>..<xupdate:otherwise..>
> > construct.
> 
> This all seems very document-centric to me.  Database operations are
> typically performed on sets of objects, not single objects, and they
> only become single objects when the participating data sets are narrowed
> into a single object through criteria.  A database that can't perform
> set operations with a single call is of little use to anyone.  The
> reason why relational databases are so powerful is because I can
> conditionally update or delete a million records with a single line of
> SQL.    I dunno, maybe I misread something here.

I don't see the problem to do this (i.e. update million records) with the
current XUpdate spec. The usage of XPath allows to select a set of objects, not
only single objects! You can select almost everything with XPath, honestly I
don't know a query that I can't express via XPath. The query would be really
complex but you can do it.

> > Also, the remainind construct in XEditor which I find useful is the
> > "path-prefix" attribute, the reason this exists is to allow the XUpdate
> > operation to take place on a specified subtree of the main document
> > specified by the path-prefix value which is prepended to the generated
> > XPaths (see http://www.openhealth.org/editor/editorgen.xsl for usage.
> > 
> > 1) this attribute is optional
> > 2) when the document to be updated does have a deep hierarchy it simplifies
> > the update list.
> > Think of it this way: suppose the entire world is in a huge XML document,
> > path-prefix allows the update operation to be carried out on a particular
> > well defined location within what is otherwise a morass.
> 
> I can't buy into the argument that the world is one big XML document. 
> Most XML Databases will provide functionality for pulling in XML from
> non-XML data sources, and the locations of these legacy data stores,
> won't fit into the XPath specification.  A path prefix can definitely
> apply to a document-rooted XPath, but if you extend the notion of what
> the prefix can contain into the realm of heterogenous data stores, you
> can no longer encapsulate that data cleanly with an XPath.

Tom, can you explain the base architecture of dbXML a little bit more. With this
information I'm hopefully able to understand some of your ideas. We had many
discussions here at SMB about 'XUpdate' that result in conflictive facts (to
your proposed approach)


Lars

-- 
___________________________________________________________________
Lars Martin                                 mailto:[EMAIL PROTECTED]
SMB GmbH                                     http://www.smb-tec.com

------------------------------------------------------------------
Post a message:          mailto:[EMAIL PROTECTED]
Unsubscribe:             mailto:[EMAIL PROTECTED]
Contact adminstrator:    mailto:[EMAIL PROTECTED]
Read archived messages:  http://www.xmldb.org/
------------------------------------------------------------------

Reply via email to