Database Reference
In-Depth Information
defined with object-oriented constructors,
OntoQL adapts and extends SQL99 provid-
ing powerful relational-object operators;
data access at the ontological level, NCCO
the SIOC namespace is set, each element without
prefix (e.g., Item) is searched in this ontology. This
query retrieves all instances of Item. To retrieve
only direct instances, one can use the ONLY op-
erator like in SQL99 (i.e., ONLY(Item)).
layer. OntoQL provides a View Definition
Language for defining and querying de-
fined concepts;
data access at the ontological level, LO
Requirement 2 (Definition of
Non Canonical Concepts)
layer. Each class and each property can be
referenced by a name (in a given natural
language) in an OntoQL query.
OntoQL provides a View Definition Language to
define non canonical concepts. Example . Create
the class PostDupont defined as all messages of
the user Dupont.
Ontologies being recorded in OBDBs, OntoQL
provides an Ontology Definition, Manipulation
and Query Language (ODL, OML and OQL) to
exploit them. To keep a uniform syntax, these
languages have been designed to keep a syntax
near SQL99. They are based on a core ontology
model that contains the main constructors of exist-
ing ontology models. This core ontology model is
an object-oriented model composed of classes and
properties named entities and attributes to distin-
guish them from ontology classes and properties.
This core ontology model can be extended with
new entities and new attributes using the ODL.
Since this model can be extended, names of enti-
ties and attributes are not encoded as keywords
in the OntoQL grammar but they are prefixed by
the character #.
CREATE #Class PostDupont AS VIEW
UNDER Post
CREATE VIEW OF PostDupont AS
SELECT * FROM Post AS p
WHERE p.has_creator.last_name =
'Dupont'
Explanation. The first statement creates the class
PostDupond as a non canonical concept (keyword
VIEW), subclass of Post. The second statement
defines the extension of this class by providing an
OntoQL query that computes its instances.
When non canonical classes are defined using
OWL constructors (e.g., restrictions), an inference
engine can automatically compute subsumption
relationships between canonical and non canonical
concepts. OntoQL (like RQL) has not this capabil-
ity: canonical and non canonical classes must be
placed manually in the hierarchy.
Requirement 1 (Queries at
the Ontological Level)
Requirement 3 (Linguistic
Exploitation)
Ontological queries can be expressed following
a SQL-like syntax. Example . Retrieve instances
of the class Item.
In OntoQL, each class and each property can be
referenced by a name in a given natural language.
It makes it possible to write the same query in
many natural languages. Example. Return the first
name and last name of users with a query written
using English names of concepts and one using
French names (Table 2).
SELECT uri
FROM Item
USING NAMESPACE 'http://rdfs.org/
sioc/ns'
Explanation . The USING clause is used to
define a default namespace for the query. When
Search WWH ::




Custom Search