xml-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Jones <jo...@nceas.ucsb.edu>
Subject Re: Saving XML to DB
Date Mon, 06 Nov 2000 23:55:54 GMT


We've been working on an XML database that we call Metacat that uses
Xerces as its XML parser.  Metacat exposes an API that allows us to
store arbitrary XML documents in a relational database (via JDBC)
without creating custom tables for each XML doctype. Right now we are
using Oracle as a backend, but other databases are possible as well
(but haven't been tested).  Metacat enables structured search of the
document contents (ie, we don't just store the XML in a BLOB) even
though the database has no prior knowledge of the XML schema/dtd.   

This is all accomplished by modeling the data as the DOM tree rather
than the explicit schema of the XML document, and then providing a
Java servlet layer on top of that.  We have a limited subset of XPath
query capability working against the document store, and have done
some clever query optimization work.  We also have enabled XSLT
transformations of the XML documents on output, so it is simple to
develop a web-based presentation of the metadata for an arbitrary
number of XML document types.

Our research project, the Knowledge Network for Biocomplexity, is
doing this work in the context of a metadata repository for a
distributed data network, but I think the effort is generically
useful.  It has many similarities to the dbXML project and the Lore
project at Stanford in its goals, but it utilizes existing relational
databases and so gains several enterprise features right out of the
box. It also has similarities to the Oracle XSQL tools that they
ship, but doesn't require a template map to convert XML to relational
form.  We find this approach substantially more flexible than the
XSQL approach, although they both have their uses (xsql is v. useful
for integrating xml data with and converting it to legacy RDBMS

Check out http://www.nceas.ucsb.edu/kdi for more information on the
project.  That's where we'll be posting our code as we make it

Unfinished features that are still on our TODO list include:
	- DOM API support (we use a simpler API now)
	- Full XPath support
	- Namespace support

If you're interested in the products, let me know and I'll see what
we can arrange.  We are currently in a testing phase, and plan our
first major release for January 2001, although I could probably
arrange for you to test our current beta version now.

Take care,

Howard Lee wrote:
> Hi folks,
> What is the best way to save XML to DB when there are no tables
> that are specific to the XML? We want to find a way to save any XML
> to DB, and able to update the XML and so on. What's the best way?
> Would XML database (e.g. excelon?) help our situation? Is there a
> plug in for Oracle to do that? Thank you very much for any
> suggestions.
> howard
> --------------------------------------------------------------------
> - In case of troubles, e-mail:     webmaster@xml.apache.org
> To unsubscribe, e-mail:          general-unsubscribe@xml.apache.org
> For additional commands, e-mail: general-help@xml.apache.org

- -- 
Matt Jones                                    jones@nceas.ucsb.edu
National Center for Ecological Analysis and Synthesis (NCEAS)

Version: PGP 6.5.1i


View raw message