W3C home > Mailing lists > Public > public-cwm-talk@w3.org > October to December 2004

Re: Cwm module to load RDBMS data

From: Graham Klyne <GK-lists@ninebynine.org>
Date: Thu, 21 Oct 2004 17:33:57 +0100
Message-Id: <5.1.0.14.2.20041021172146.031273e8@127.0.0.1>
To: "Jones, David H" <david.h.jones@boeing.com>, <public-cwm-talk@w3.org>

I think it's a grand idea!

I did something vaguely related for my Swish software [1] to import data 
from a (Excel) spreadsheet, via CSV files, which I subsequently used to 
create an RDF-driven test suite for the new URI specification.  By adding 
annotations to the spreadsheet, or by prepending the CSV file with 
equivalent information, the URIs used for (column) properties, and RDF 
datatypes to be used for values in a column, can be specified.

A sample CSV file is at:
   http://www.ninebynine.org/Software/HaskellUtils/Network/UriTest.csv

Other related files can be found in the same directory:
   http://www.ninebynine.org/Software/HaskellUtils/Network/

The main CSV parsing module for Swish is here:
   http://www.ninebynine.org/Software/HaskellRDF/RDF/Graph/CSVParser.hs
Test module:
   http://www.ninebynine.org/Software/HaskellRDF/RDF/Graph/CSVParserTest.hs

This feature is not yet in a packaged release of Swish, so you'd need to 
poke around the Haskell source files to get all the pieces of a working 
Swish with this capability:
   http://www.ninebynine.org/Software/HaskellRDF/RDF/
   ... and 4 subdirectories
   http://www.ninebynine.org/Software/HaskellUtils/
   ... with subdirectory Network

[1] General description of Swish:
   http://www.ninebynine.org/RDFNotes/Swish/Intro.html

#g
--

At 08:15 21/10/04 -0700, Jones, David H wrote:
>I was thinking that it would be very interesting and useful to be have a 
>cwm and Euler module that loaded RDBMS data into the inference engine for 
>subsequent processing.
>
>My particular preference would be to do this in a very generic way, 
>preserving some of the table column/table information/structure.  The 
>advantage of this minimal processing is that publishing data from the DB 
>is made as simple as possible.  Subsequent processing with rules could 
>transform/integrate the data to conform with other ontologies.
>
>One advantage of such a module would be to facilitate publishing of RDBMS 
>data in RDF.  Another would be to facilitate combining RDBMS data with 
>other data and rules.
>
>Attached are several n3 files that illustrate one approach.
>
>I'm interested in stimulating a discussion to see what others think of the 
>general idea, to have others help me refine the approach I've outlined, or 
>to propose alternate approaches.  Assuming some sort of consensus could be 
>reached, I'd be willing to do a trial implementation for cwm of such a module.
>  <<rdb.n3>>  <<people.n3>>
>Regards,
>
>David H. Jones
>Boeing Phantom Works,
>Mathematics & Computing Technology
>425-865-6924
>425-865-2964 (FAX)
>david.h.jones@boeing.com
>
>
>

------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact
Received on Thursday, 21 October 2004 16:41:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:01 GMT