RDFS and OWL Reasoning for Linked Data - Reasoning Web - page 102

Given the data in Fig. 2, this pattern (and thus the query) would obtain
the following solutions (writing solution mappings μ in tabular form):
?X ?L ?R
dbr:SAP AG "SAP AG"@en 1.622E10
dbr:Siemens "Siemens"@de 7.829E10
Notably, the revenue for IBM is not returned (although it could be calcu-
lated from the EUR-USD exchange rate).
The next query, on the contrary, asks for the date of the latest article
published about each element in the SKOS scheme nytimes:nytd_org (i.e.,
a different way of asking for “organisations”, but using the NYT RDF
vocabulary):
Query 2
SELECT ?X ?D
WHERE { ?X skos:inScheme nytimes:nytd_org. ?X nytimes:latest_use ?D .}
with the following solutions:
?X ?D
nyt:75293219995342479362 2010-05-13
nyt:N82918236209763785922 2009-11-06
nyt:49586210195898795812 2010-04-27
Query 1 and Query 2 could each be answered by staying within a single
site (that is, Query 1 only would obtain answers from data at dbpedia.org ,
whereas Query 2 would only produce answers with the data at data.nytimes.o
respectively) and - at least for our sample - answers for either query can
be obtained from the individual dataset. However, the real power of Linked
Data lies in combining data from multiple datasets to obtain answers over
their combined content. For example Query 3 combines knowledge from
both sites and asks for the latest NYT article dates of IBM (using its NYT
identifier) and its revenue in USD:
Query 3
SELECT ?X ?D ?R
WHERE { nyt:49586210195898795812 nytimes:latest_use ?D .
nyt:49586210195898795812 owl:sameAs ?X .
?X dbo:revenueUSD ?R .}
Again assuming the entire graph of Fig. 2 as input, this query would obtain
the single result
?X ?D ?R
dbr:IBM 2010-04-27 1.06916E11
As a further example, let Query 4 ask for all foaf:Agent s .