1. SPARQL in the Semantic Web
Jan Carl Beeck Pepper
1

2. Agenda
• Definitions and Motivation
• SPARQL
2

3. Definitions and Motivation
• Statement (or triple):
– Small piece of knowledge (a single fact).
– It has Subject-Predicate-Object.
– Ex. Evidence A is_a physiotherapy evidence.
– Ex. <subj0> <pred0> <obj0>
• Subject (resource) and Object (value):
– Names for things in the world.
• Predicate (property):
– Name of a relation that connects two things.
3

4. Definitions and Motivation
• Semantic Web:
– It is built on top of the current Web.
– Besides the HTML constructs, it contains some
“statements” that can be collected by an agent.
– The agent organizes and connects the statements
into a graph format (data integration).
– Automatic data integration on the Web can be
powerful and can help a lot when it comes to
information discovery and retrieval.
4

5. Definitions and Motivation
• Query-based-language:
– The agent should be able to process some
common queries that are submitted against the
statements it has collected. After all, without
providing a query interface, the collected
statements will not be of too much use to us.
5

6. Definitions and Motivation
• Linked Data:
– A collection of machine-understandable
statements, published without having them
related to any Web site at all.
• Web of Data:
– Interchangeable terms for the Semantic Web.
6

7. Definitions and Motivation
• Resource description framework (RDF):
– The building block for the Semantic Web.
– Standard for encoding metadata.
• Metadata: describe the data contained on the Web.
• Machine understandable (also interoperability).
• Domain independent.
– Describe any resources and their relations existing
in the real world.
– RDF is for the Semantic Web what HTML has been
for the Web.
7

8. Definitions and Motivation
• RDF Schema (RDFS):
– Stands for RDF Schema.
– Common language, or, a vocabulary, where
classes, sub-classes, properties, and also relation
between the classes and properties are defined.
– Domain-specific.
– Allow the creation of distributed RDF documents.
8

9. Definitions and Motivation
• Web Ontology Language (OWL):
– Is the most popular language to use when creating
ontologies.
– Is build upon RDF Schema.
– Has the same purpose as RDF Schema.
• Classes, properties, and their relationships for a specific
application domain.
– Provides the capability to express much more
complex and richer relationships (better
expressiveness).
9

10. Definitions and Motivation
• Web Ontology Language (OWL):
– Axiom: basic statement (basic piece of
knowledge).
– A collection of axioms is an OWL Ontology.
– Protégé is free OWL editor.
• IRI:
– Stands for Internationalized Resource Identifiers
(like URIs with Unicode characters).
10

11. Definitions and Motivation
• Computer Ontology:
– Reflects the structure of the world.
– Is often about structure of the concepts.
– Each statement collected by an agent represents a
piece of knowledge. Therefore, there has to be a
way (a model) to represent knowledge on the
Web. Furthermore, this model of representing
knowledge has to be easily and readily processed
(understood) by machines.
11

12. Definitions and Motivation
• Computer Ontology:
– An application can understand a given ontology;
that means the application can parse the ontology
and create a list of axioms based on the ontology,
and all the facts are expressed as RDF statements.
12

13. Agenda
• Definitions and Motivation
• SPARQL
13

14. SPARQL
• SPARQL: Querying the Semantic Web.
– Pronounced “splarkle”
– Stands for SPARQL Protocol and RDF Query
Language.
– Locate specific information on the machine-
readable Web.
– The Web can be viewed as a gigantic database.
14

15. SPARQL (cont)
• Related concepts:
– RDF data store: is a special database system built
for the storage and retrieval of RDF statements.
• Every record is a short statement in the form of
subject-predicate-object.
• Store RDF statements and retrieve them by using a
query language.
– Triple pattern: any or all the subject, predicate,
and object values can be a variable.
• <http://danbri.org/foaf.rdf#danbri> foaf:name ?name.
15

16. SPARQL (cont)
• Graph pattern: is used to select triples from a
given RDF graph.
– Is a collection of triple patterns.
– { ?who foaf:name ?name.
?who foaf:interest ?interest.
?who foaf:knows ?others. }
– Note: FOAF (friend of a friend) is an Ontology (a
group of properties that describes a person) and a
collection of RDF statements.
16

17. SPARQL (cont)
• SPARQL engine:
– Tries to match the triples contained in the graph
patterns against the RDF graph, which is a
collection of triples.
– Once a match is successful, it will bind the graph
pattern’s variables to the graph’s nodes, and one
such variable binding is called a query solution.
17

18. SPARQL (cont)
• SPARQL endpoint:
– Interface that users (human or apps) can access to
query an RDF data store by using SPARQL query
language.
• Web-based application.
• Set of APIs that can be used by an agent.
– Ex. Joseki Web-based SPARQL.
• http://sparql.org/sparql.html
18

19. SPARQL (cont)
• SPARQL Query Language:
– SELECT query (most frequently used).
– ASK query.
– DESCRIBE query.
– CONSTRUCT query.
19

20. SPARQL (cont)
• Structure of a SELECT Query:
– # base directive
BASE <URI>
# list of prefixes
PREFIX pref: <URI>
...
# result description
SELECT...
# graph to search
FROM . . .
# query pattern
WHERE {
...
}
# query modifiers
ORDER BY... 20

21. SPARQL (cont)
• Ex. Find all the picture formats used by Dan Brickley’s friends (from graph
http://danbri.org/foaf.rdf#danbri).
21

22. SPARQL (cont)
• The query finds all the picture format used by Dan Bricley's friends.
• Base define the source file (graph) which link is: http...
• prefix define the ontology of persons foaf which link is: http...
• the other prefix define the image format ontology dc which link is: http...
• select * from the source graph
• where is defined the term dambri in the foaf ontology throw the knows
attribute and store that information in the variable ?friend
• where ?friend has a description of the image throw the attribute
foaf:depiction and store that information in the variable ?picture
• where ?picture has the name of the image format throw attribute
dc:format and store the name in the variable ?imageFormat
22

23. SPARQL (cont)
23

24. SPARQL (cont)
• Optional keyword: is needed because RDF
data graph is only a semi-structured data
model.
– i.e. two instances of the same class type in a given
RDF graph may have different set of property
instances created for each one of them.
– The query says, find all the people known by Dan
Brickley and show their name, e-mail, and home
page information if any of this information is
available.
24

25. SPARQL (cont)
25

26. SPARQL (cont)
26

27. SPARQL (cont)
• Solution modifiers:
– Distinct: eliminate duplicate solutions from the
result.
– Order by:
• Asc (): ascending.
• Desc (): descending.
– Limit: set the maximum number of solutions.
– Offset: sets the number of solutions to be skipped.
27

28. SPARQL (cont)
• Filter keyword to add value constraints,
functions and operators.
28

29. SPARQL (cont)
• Union keyword:
– A query expressed by multiple graph patterns that
are mutually exclusive, and any solution will have
to match exactly one of these patterns
(alternative match).
29

30. SPARQL (cont)
• Multiple graphs:
– SPARQL allows us to query any number of named
graphs.
30

31. SPARQL (cont)
• Construct query:
– Returns a new RDF graph.
• Describe query:
– Return an RDF graph whose statement are
determined by the query processor.
• Ask query:
– The query processor simply returns a true or false
value.
31

32. SPARQL (cont)
• Aggregate functions:
– COUNT
– SUM
– MIN/MAX
– AVG
– GROUP_CONCAT
– SAMPLE
32

33. SPARQL (cont)
• Other operators and functions:
– NOT EXISTS
– MINUS
– Concat() : for expressions in a query.
– INSERT DATA
– DELETE DATA
– CREATE [SILENT] GRAPH <uri>
– DROP [SILENT] GRAPH <uri>
33

34. References
• Liyang Yu (2011). A Developer’s Guide to the
Semantic Web. Springer. ISBN: 978-3-642-
15969-5.
• (2011) Huang J, Abadi D.J. and Ren K. Scalable
SPARQL Querying of Large RDF Graphs.
34