* source: sparql-aderis - Revision 143: /trunk

svn checkout sparql-aderis-read-only

* execute:


 java -jar ADERIS.jar,,,, "select ?s1 ?o1 ?o2 where{ ?s1 <>   <>.
?s1   <> ?o1. ?o1   <> ?o2. <>  
 <> ?o2.}"

Note: Use of "Prefix" tends to produce problems, such as " ?drug_drugbank drugbank:category drugbank:Anti-Allergic-Agents." in query 4 and " ?drug_drugbank drugbank:category drugbank:Anticonvulsants." in query 5, the objects with Predix will take errors. Query 4 should be written as "?drug_drugbank drugbank:category <>."


query1: results:zero.

Query\time(ms) time 1 time 2 time 3 time 4time 5 avg
case1 27010252492571025481 2614725647

* Analysis:

1. When doing queries, ADERIS firstly queries each data sources using SPARQL to find out the
  number of distinct predicate values(Data about subjects/objects  not retrieved.)

  select distinct ?p where { ?s ?p ?o };
  select count(?p) where { ?s ?p ?o };

'''potential consequences:'''

 Without using caches and any metadata information pre-fetched, the system query each endpoint in real time,
which could take a long time, especially, when the number of endpoints increases.