W3C home > Mailing lists > Public > public-xg-rdb2rdf@w3.org > May 2008

Re: rdfs rdb vocabulary (again)

From: Juan Sequeda <jsequeda@cs.utexas.edu>
Date: Mon, 26 May 2008 10:43:04 -0600
Message-ID: <f914914c0805260943r738641bbu926aff614cc04712@mail.gmail.com>
To: "Paul Tyson" <phtyson@sbcglobal.net>
Cc: "public-xg-rdb2rdf@w3.org >> public-xg-rdb2rdf" <public-xg-rdb2rdf@w3.org>
Hi Paul and All

We share the same approach which was our position paper [1] and denominate
it as "direct mapping"

In the State of the Art[2], the paper by Man Li exposes this idea. We have
given it a more formal approach and a notion of completeness here [3] (it
has not been published yet, therefore not referencable). A survey paper on
"direct mapping" is coming soon also.

We are currently developing a system that follows this approach by
automatically creating an ontology out of the sql schema utilizing sql-ddl
and our next step is to translate sparql to sql queries on the fly.

[1] http://www.w3.org/2007/03/RdfRDB/papers/sequeda.pdf
[2] http://esw.w3.org/topic/Rdb2RdfXG/StateOfTheArt
[3] http://www.cs.utexas.edu/~jsequeda/pub/sql2sw.pdf
On Mon, May 26, 2008 at 8:31 AM, Paul Tyson <phtyson@sbcglobal.net> wrote:

>
> I am not a member of the XG, but have been following the discussions. It is
> apparent that there are any number of ways to expose RDB data as RDF, but
> not so apparent that there are any aspects of these approaches that could or
> should be standardized.
>
> I suggested in an earlier note (
> http://lists.w3.org/Archives/Public/public-xg-rdb2rdf/2008Mar/0029.html)
> that one aspect of standardization that would advance the state of the art
> is an RDFS or OWL formulation of the relational model itself.  Any method of
> getting RDB data to RDF has to deal with the relational model either
> explicitly or implicitly.  By making an explicit model of RDB in RDF, we at
> least provide a reference point for all implementations.
>
> Here is a rough cut straw man proposal. It has several defects, and raises
> many questions.  I would like to submit this to the XG for discussion on at
> least 3 topics:
>        1) the validity and usefulness of the schema itself (or any such
> schema);
>        2) possible use in automated SQL-SPARQL rewriting (2-way);
>        3) possible use as an intermediate format between RDB and
> domain-specific ontologies.
>
> @prefix rdfs: http://www.w3.org/2000/01/rdf-schema# .
> @prefix rdf: http://www.w3.org/1999/02/22-rdf-syntax-ns# .
> @prefix rdb: http://www.w3.org/xg-rdb2rdf/rdb# .
>
> rdb:Relation a rdfs:Class . # a database table
> rdb:relationName a rdf:Property ; # the table name
>        rdfs:domain rdb:Relation .
> rdb:header a rdf:Property ;
>        rdfs:domain rdb:Relation ;
>        rdfs:range rdb:RelationHeader .
> rdb:body a rdf:Property ;
>        rdfs:domain rdb:Relation ;
>        rdfs:range rdb:RelationBody .
> rdb:RelationHeader a rdfs:Bag . # of TypeDefinitions
> rdb:TypeDefinition a rdfs:Class .
> rdb:typeName a rdf:Property ; # of a TypeDefinition or TypedValue
>        rdfs:range rdb:TypeDefinition ;
>        rdfs:range rdb:TypedValue .
> rdb:underlyingType a rdf:Property ; # of a TypeDefinition
>        rdfs:range rdb:TypeDefinition .
> rdb:primaryKey a rdf:Property ; # boolean
>        rdfs:range rdb:TypeDefinition .
> rdb:foreignKey a rdf:Property ; # reference to foreign typeName
>        rdfs:range rdb:TypeDefinition .
> rdb:RelationBody a rdfs:Bag . # of Tuples
> rdb:Tuple a rdfs:Bag . # of TypedValues
> rdb:TypedValue a rdfs:Class .
> rdb:value a rdf:Property ; # of a TypedValue
>        rdfs:range rdb:TypedValue;
>
>


-- 
Juan Sequeda

Research Assistant - Ph.D Student
Dept. of Computer Sciences
The University of Texas at Austin
http://www.cs.utexas.edu/~jsequeda
jsequeda@cs.utexas.edu

Semantic Web in Austin: http://juansequeda.blogspot.com/
Received on Monday, 26 May 2008 16:43:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 26 May 2008 16:43:46 GMT