- From: Eric Prud'hommeaux <eric@w3.org>
- Date: Thu, 22 Apr 2010 08:05:07 -0400
- To: Georgi Kobilarov <georgi.kobilarov@uberblic.com>
- Cc: Michael Hausenblas <michael.hausenblas@deri.org>, public-rdb2rdf-wg@w3.org
Hi Georgi, we're working on a Use Cases and Requirements document a mapping language from relational data to RDF graphs. An obvious use of this is to define a virtual RDF graph for a given RDB and transform a SPARQL query over that graph into a SQL query over the relational data. About an hour ago, I had hoped to find the SQL table structure which backed Wikipedia infoboxes. I believe I have since learned that the structured data in infoboxes is microparsed from blobs containing the page text. This is addressed by microparsing http://www.w3.org/2001/sw/rdb2rdf/use-cases/#reqs-functional-5 but if the parsing is complicated, the functionality is unlikely to be standardised. I have the vague impression that uberblic does query transformation à la http://www.w3.org/2001/sw/rdb2rdf/use-cases/#Queries_Over_the_RDF_Graph. to http://www.w3.org/2001/sw/rdb2rdf/use-cases/#RDF_Graphs_as_Non-materialized_Views Is that so? Is there a budding SQL structure for infoboxes which we should include in our use cases? Note, cc-ing public-rdb2rdf-wg in a reply will require a dance through the W3C Archive Approval system <http://www.w3.org/2002/09/aa/> . -- -ericP
Received on Thursday, 22 April 2010 12:05:42 UTC