[
Lists Home |
Date Index |
Thread Index
]
I don't know if I agree with that assessment, Joseph, at least regarding
relational tables. I haven't done much implementation work myself, other
than munging my own XQuery engine to provide a native internal format for
RDF triples, but my impression is that it's not that difficult to implement
a backend store for RDF in a relational table. Each triple ultimately just
boils down to three integer columns, which are either pointers into tables
of indexed URIs or literal data instances. I've been looking at doing
exactly that with my own engine: implementing an optional SQL-based backend
as a user option, and at first glance it doesn't look that difficult.
Does anybody on the list have contrary experience? I know at least a few
members of the RDF Data Access Working Group have implemented backends in
SQL. I was quite intrigued to look over a few shoulders at the last
face-to-face meeting of the group and watch people hacking away on big
repositories in MySQL.
I agree that implementing via a native XML database doesn't buy you much,
other that the ability to easily suck in RDF that's been stored as RDF/XML
and to serialize to the same format on output. That's definitely extremely
useful for interop. However there's too big an impedance mismatch between
the sort of things you'd like to do with XML and with RDF to buy you much of
value on the query side.
Best,
Howard
> -----Original Message-----
> From: Joshua Allen [mailto:joshuaa@microsoft.com]
> Sent: Monday, May 17, 2004 11:12 AM
> To: John Song; Xml-dev
> Subject: RE: [xml-dev] RDF storage
>
>
> > and the arising problems? the most difference betwwen RDF
> > and pure XML is the data model which is hierarchical tree for XML
> > and graph for RDF.so if storing RDF in NXD,maybe it is
> > needed to redesign the storage model and also mapping rdf query
> > like RQL into XML query language like XQuery.
> > It seems that the most appealing thing to store RDF in NXD
> > is that i can utilize the index and other optimizing
>
> You are correct; the difference in data model makes it difficult to get
> any advantage from Xquery. The same applies to indexing though. The
> idices which are supported by native XML databases are not RDF-aware.
> However, we could make the same criticism of relational databases -- the
> RDF data model is not at all like the relational data model; so vendors
> have been forced to implement custom indexing and query schemes atop
> relational engines. Since it requires a lot of custom work either way,
> I do not think that either relational or native-XML have a distinct
> advantage over the other, except for the fact that relational engines
> are more mature, and targeting relational makes it easier to port a RDF
> engine to multiple platforms. A native XML database would perhaps have
> an easier time preserving the full fidelity of the input RDF documents,
> but it is also possible for a relational-backed engine to store the
> source RDF in a CLOB when shredding the RDF to relational tables.
>
> -----------------------------------------------------------------
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
>
> The list archives are at http://lists.xml.org/archives/xml-dev/
>
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://www.oasis-open.org/mlmanage/index.php>
>
|
