On 4/5/11 3:42 PM, William Waites wrote:
So I don't have answers to your questions, but do have some
observations about the results, particularly the counts of
distinct predicates.
The top one is rdf:type which makes sense. Below that we
have ones used in reification. Who knew there was actually
that much reified data out there? I wonder where this comes
from and what about the consensus that this is not a good
idea and should be deprecated?
SELECT DISTINCT ?graph, COUNT(?s) AS ?count WHERE {
GRAPH ?graph { ?s
?p<http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement> }
} ORDER BY DESC(?count) LIMIT 50
This query times out, but it would be interesting to know
the answer, who is the source of all of these reifications?
Yes, that will timeout via the public SPARQL endpoint. We'll run it
internally to get the numbers.
Next is rdfs:label, ok, fine. After that, a sizeable chunk
of data has to do with rows and columns in CSV tables that
comes from data.gov.
No, that's RDF from RPI's (Jim Hendler's team) conversion of Data.Gov
datasets. That accounts for about 6.4 Billion triples re. total
contribution.
How is a mechanical transliteration
from CSV to RDF without any modelling useful?
That's a question for the team at RPI :-)
It just makes
the data a couple of orders of magnitude bigger and a few
more orders of magnitude more cumbersome to deal with.
Yes and No. As will all of these matter utility lies in the eyes and
fingers of the data beholder.
I
mean, being able to refer to a specific spreadsheet cell is
useful but how does actually materialising all of them do
anything but take up disk space and slow down queries?
See comments above :-)
Cheers,
-w
--
Regards,
Kingsley Idehen
President& CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen
--
Regards,
Kingsley Idehen
President& CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen
