Make Databases accessible for the Semantic Web

Besides (re)creating data manually, automatic generation or extraction of information already there is a possibility. One more specific possibility are databases which already contain huge amount of highly structured data (for example content management systems or the so called deep web). But how can we make this data part of the semantic web?
1 answer

Make Databases accessible for the Semantic Web - Solution Proposal

First it's very important to find out how much structure needs to be maintained.

A first approach could be:

  • find a uri for every row of every table (will be called <tuple> from now on).
  • For every column and for every row: Create an RDF triple:
    <tuple> <columnname> <value> .

But if also relations between tables has to be "transported" to RDF, this concept has to be implemented in the parser. The first approach would be to point to other entities with the RDF built in predicates. Use rdf:Bag for 1:n and n:m relations.

Another concept which isn't consired in this schema are constraints like unique columns. to implement this, a new vocabulary has to be defined.

For more information see G. Lausen, M. Meier, M. Schmidt: SPARQLing Constraints for RDF in Proceedings of the 11th international conference on Extending database technology: Advances in database technology, 2008, 499-509.