OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: Storage of XML documents & Learning from history

[ Lists Home | Date Index | Thread Index ]
  • From: Michael Rys <mrys@microsoft.com>
  • To: "'Kay Michael'" <Michael.Kay@icl.com>, "'Dylan Walsh'" <Dylan.Walsh@Kadius.com>, xml-dev@XML.ORG
  • Date: Fri, 12 May 2000 10:37:42 -0700

> > I problem I am looking at is how best to store XML documents, 
> > an area I'm sure a lot of developers are hitting right now.
> The principal decision is between fine-grained storage (in which the
> document structure is known to the DBMS) and coarse-grained 
> storage (in which it isn't).
> Fine-grained storage is necessary if you want to use DBMS  queries to
> the information, but it is incredibly expensive, regardless  of whether
> use a relational or an object DB. Our experience is that fine-grained
> storage is very rarely necessary or cost-effective: just extract the few
> pieces of information needed to do queries, store these in separate
> and hold the rest of the XML document as a blob.

I agree.

> If you do need fine-grained storage, use an object database; relational
> storage of these structures will in general be grotesque. 

Here I disagree. This depends heavily on what you want to do with the data
in the document. Assuming that the fully idempotent recomposition of the
document is of lower importance (otherwise, store the coarse-grained views
as well as mentioned below) and instead  the main reason for fine-grained
storage is to combine and join the data with other data,  relational storage
may be as adequate.

Shameless plug:
For a tool in a relational context to do both look at the OpenXML XML rowset
provider in SQLServer 2000 (available as beta at
http://msdn.microsoft.com/sql). I assume other database vendors have similar

> Even with an
> object DB, consider storing both the fine-grained and the coarse-grained
> views, so that you never have to reassemble the full document 
> by finding all its bits.

One of the hard issues here is of course to preserve integrity accross these

> The old analogy here is that of taking your car to pieces 
> every time you
> want to park it overnight...

You don't do that :-)

Best regards

This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS