You likely want to use a sub-query to restrict to the documents matching the FILTER
and then select the rest of the stuff you are actually interesting in e.g.
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX dc : <http://purl.org/dc/elements/1.1/>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
SELECT ?document ?subject ?title ?description ?language
WHERE
{
{
SELECT ?document
WHERE
{
?document dc:subject ?subject .
FILTER(REGEX(?subject, REPLACE("search term", " ", "|"), "i"))
}
}
?document dc:title ?title ;
dc:description ?description ;
dc:subject ?subject ;
dc:language ?language .
}
Note that this is still going to give you a row for each document-subject combination so if you have a document with 3 subjects you'll get three rows for that document. If you want to combine documents into a single row then you can use GROUP BY
and then a GROUP_CONCAT
aggregate, there are other questions already on Stack Overflow that detail how to do that.
Notes
Also note that using simple string concatenation to inject constants into your query is ill advised, take a look at Jena's ParameterizedSparqlString for a more user friendly and SPARQL injection proof API for building queries.