Data Warehouses and the Semantic Web - Data Warehouse Systems: Design and Implementation - page 538

SELECT ?firstName ?lastName
WHERE { ?emp a ex:Employee ; ex:Employee#FirstName ?firstName ;

ex:Employee#LastName ?lastName ; ex:Employee#HireDate ?hireDate .
FILTER( ?hireDate > = '' 1992-01-01 '' ∧∧ xsd:date &&
?hireDate < = '' 1994-12-31 '' ∧∧ xsd:date) }

Filter conditions are Boolean expressions constructed using the logical
connectives && (and), (or), and ! (not).

The FILTER keyword can be combined with the NOT EXISTS keyword to
test the absence of a pattern. For example, the query “First name and last
name of employees without supervisor” reads in SPARQL as follows:

||

SELECT ?firstName ?lastName
WHERE { ?emp a ex:Employee ; ex:Employee#FirstName ?firstName ;

ex:Employee#LastName ?lastName .
FILTER NOT EXISTS { ?emp ex:Employee#Supervisor ?sup . }}

The OPTIONAL keyword is used to specify a graph pattern for which the
values will be shown if they are found. For example, the query “First and
last name of employees, along with the first and last name of her supervisor,
if she has one” can be written in SPARQL as follows:

SELECT ?empFirstName ?empLastName ?supFirstName ?supLastName
WHERE { ?emp a ex:Employee ; ex:Employee#FirstName ?empFirstName ;

ex:Employee#LastName ?empLastName .
OPTIONAL ?emp ex:Employee#Supervisor ?sup .

?sup a ex:Employee ; ex:Employee#FirstName ?supFirstName ;
ex:Employee#LastName ?supLastName . }}

{

Notice that the OPTIONAL keyword behaves in a way similar to an outer
join in SQL.

Aggregation and Sorting in SPARQL

Aggregate functions summarize information from multiple triple patterns
into a single one. SPARQL provides the usual aggregate functions COUNT ,
SUM , MAX , MIN ,and AVG . In addition, along the lines of SQL, before
summarization, triples may be grouped using the GROUP BY keyword, and
then the aggregate function is applied to every group. Furthermore, filtering of
groups may also be performed with the HAVING keyword, like it is done with
the FILTER clause for ungrouped sets. Finally, the result can be sorted with
the ORDER BY clause, where every attribute in the list can be ordered either
in ascending or descending order by specifying ASC or DESC , respectively.
Consider the query “Total number of orders handled by each employee, in
descending order of number of orders. Only list employees that handled more
than 100 orders.” This query is expressed in SPARQL as follows: