W3C home > Mailing lists > Public > public-xg-rdb2rdf@w3.org > January 2009

Survey of Current Approaches for mapping Relational Databases to RDF Comment...

From: Ezzat, Ahmed <Ahmed.Ezzat@hp.com>
Date: Sat, 3 Jan 2009 04:51:04 +0000
To: "public-xg-rdb2rdf@w3.org" <public-xg-rdb2rdf@w3.org>
Message-ID: <3B7AE9BA67C72B4891EF21842246A21C4049841E9F@GVW1097EXB.americas.hpqcorp.net>


Good work.  I found two versions of the same paper, they are close but there are differences.  One is dated 11/1/08 and the other do not have date.   My comments is on segments that I think is common between the 2 versions:

1.      It is not clear why Section 2 is not being part of Section 3?

AKE> I suggest to combine and restructure as follows:

RDB to RDF Mapping Dimensions Survey:
*       Mapping Approaches:
*       Local Ontology Mapping: this approach typically is an automatic mapping of RDBMS table to rdfs:Class or to owl:Class with columns as predicate or properties of that class.  This approach provides semantic web application a means to querying a database where the application semantics is defined in terms of the database schema.
Discussion: differences between RDFS and OWL as a means for the local Ontology mapping...
*       Domain Ontology Mapping: this approach typically maps the RDBMS schema into application-specific domain ontology directly or maps the above local ontology to the same application-specific domain ontology. This approach typically will be semi-automatic.
*       Data Integration: If there are multiple data sources involved, you will need to use the Domain Ontology Mapping approach.  In addition, you will need to reconcile the semantics between different data sources or their corresponding local ontologies.  This may be semi-automatic and most probably will involve rules engine (SWRL) to reconcile semantics between the different local ontologies.
*       Mapping types: either static/ETL approach or dynamic/on-demand approach.

*       Mapping algorithms Representation: XSLT/XPath, R2O, etc..

*       Query Rewrite: generate SQL from SPARQL. It would be useful to present alternative known approaches to do this mapping.

*       Mixing SPARQL and SQL: this is like Oracle approach and I think Virtuoso...

2.      Suggest to change Section 4 Header to be homogenous with the above Section:

RDB to RDF Mapping Implementations Survey:

*       Schema mapping.
*       ...
*       ...

I like Table-1 summarizing what is out there.  It might makes sense to add another column [commercial/Open Source] and list some of the open source tools out there like Jena and others.


Ahmed K. Ezzat, Ph.D.
HP Fellow, Business Intelligence Software Division
Hewlett-Packard Corporation
11000 Wolf Road, Bldg 42 Upper, MS 4502, Cupertino, CA 95014-0691
Office: Email: Ahmed.Ezzat@hp.com Tel: 408-447-6380 Fax: 1408796-5427 Cell: 408-504-2603
Personal: Email: AhmedEzzat@aol.com Tel: 408-253-5062 Fax: 408-253-6271
Received on Saturday, 3 January 2009 04:52:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:51:39 UTC