W3C home > Mailing lists > Public > semantic-web@w3.org > April 2016

Re: RDB structure to RDF mapping

From: Pat Hayes <phayes@ihmc.us>
Date: Sat, 23 Apr 2016 23:23:43 -0500
Cc: Paul Tyson <phtyson@sbcglobal.net>, semantic-web@w3.org
Message-Id: <6BB68188-CB47-48CC-8DE4-B4FBB171CA6B@ihmc.us>
To: Alessandro Seganti <a.seganti@cognitum.eu>

On Apr 22, 2016, at 2:47 AM, Alessandro Seganti <a.seganti@cognitum.eu> wrote:

> Thank you all for your answers!
> So if I understood correctly the answer is: there is no standard to deal with modeling of the structure of RDBs. I started doing this by using instances as it seems to me the simplest way. I am trying to make an ontology as much as possible compatible with R2RML but there is not much to use apart from logicalTable and a couple of relations.
> 
> This said, do you think that is there a reason why none has thought of a standard for this kind of mapping?

It is an issue that many people have felt should be tackled, and it seems obvious how to start, but the details become very hard and have defeated all attempts to standardize such a mapping. For example, nothing in RDF can behave in the way that null behaves in RDBs, and RDF is agressively open-world in its underlying semantic model. 

I wish you well in your attempt to do this: if you feel you are succeeding, please publicize your efforts as they may be widely useful. 

Pat Hayes

> Maybe because in most use cases people want to map only specific queries to RDF graphs instead of mapping the whole database structure to an RDF graph?
> 
> Regards,
> Alessandro
> 
> Alessandro Seganti Ph.D.
> 
> Cognitum Sp. z o.o., Data Engineer
> 
> mail: a.seganti@cognitum.eu, tel: +48 22 250 2541, fax: +48 22 250 28 98
> 
> web: http://www.cognitum.eu/, fb: facebook.com/cognitum.eu
> 
> 
> On Tue, Apr 19, 2016 at 8:13 PM, Paul Tyson <phtyson@sbcglobal.net> wrote:
> On Mon, 2016-04-18 at 17:25 +0200, Alessandro Seganti wrote:
> > Hi everyone,
> > I want to formalize in RDF the table structure of a relational
> > database. What I would like to do is to formalize the links that
> > exists between all the tables inside a given relational database. This
> > would result in a graph where the nodes are the tables and the links
> > are the columns that can be used to join these tables.
> >
> >
> > My question is: is there some standard/ontology to describe this?
> >
> >
> > To my understanding the R2RML ontology is to describe mappings between
> > RDB and RDF (no RDB structure is mentioned). I also found a document
> > about direct mapping but no ontology is defined in this document.
> >
> 
> When the RDB2RDF WG was working I suggested including an RDB ontology.
> But it was considered out of scope so nothing of the sort was included
> in their final deliverables.
> 
> http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2010Oct/0055.html
> 
> My work didn't (and doesn't) involve RDBMS deeply enough to pursue this,
> so I don't know how valuable this would be for practical tasks.
> 
> Regards,
> --Paul
> >
> >
> >
> > Thank you for your help.
> > Alessandro Seganti Ph.D.
> >
> > Cognitum Sp. z o.o., Data Engineer
> >
> > mail: a.seganti@cognitum.eu, tel: +48 22 250 2541, fax: +48 22 250 28
> > 98
> >
> > web: http://www.cognitum.eu/, fb: facebook.com/cognitum.eu
> >
> >
> 
> 
> 

------------------------------------------------------------
IHMC                                     (850)434 8903 home
40 South Alcaniz St.            (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile (preferred)
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Sunday, 24 April 2016 04:24:56 UTC

This archive was generated by hypermail 2.3.1 : Sunday, 24 April 2016 04:25:00 UTC