rdfsparqlsesameopenrdf

ORDER BY solution modifier in CONSTRUCT pattern of SPARQL


I followed SPARQL 1.1 section 10.2 CONSTRUCT and have the following SPARQL query which returns all triples whose subject has type Homework and which also have an event date.

CONSTRUCT  { ?s ?p ?o } WHERE 
{
 GRAPH ?g { ?s ?p ?o } .
 { ?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>   
    <http://lod.isi.edu/ontology/syllabus/Homework> 
 }.

 { ?s <http://lod.isi.edu/ontology/syllabus/hasEventDate> ?date}.
}

Now, I want to retrieve the results in ascending/descending order of the event date, ?date. I tried adding the order by modifier as shown below, but there are parse errors in the OpenRDF workbench.

CONSTRUCT  { ?s ?p ?o } WHERE 
{
 GRAPH ?g { ?s ?p ?o } .
 { ?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>   
    <http://lod.isi.edu/ontology/syllabus/Homework> 
 }.

 { ?s <http://lod.isi.edu/ontology/syllabus/hasEventDate> ?date}.
 ORDER BY DESC(?date)

}
Encountered " "order" "ORDER "" at line 9, column 6. Was expecting one of:
"(" ... "{" ... "}" ... "[" ... <NIL> ... <ANON> ... "optional" ...
"graph" ... "minus" ... "filter" ... "true" ... "false" ... "bind" ...
"service" ... "values" ... <Q_IRI_REF> ... <PNAME_NS> ... <PNAME_LN> ...
<BLANK_NODE_LABEL> ... <VAR1> ... <VAR2> ... <INTEGER> ... <INTEGER_POSITIVE> ...
<INTEGER_NEGATIVE> ... <DECIMAL> ... <DECIMAL_POSITIVE> ...
<DECIMAL_NEGATIVE> ... <DOUBLE> ... <DOUBLE_POSITIVE> ...
<DOUBLE_NEGATIVE> ... <STRING_LITERAL1> ... <STRING_LITERAL2> ...
<STRING_LITERAL_LONG1> ... <STRING_LITERAL_LONG2> ...

Solution

  • The problem is that your ORDER BY clause is not in the correct place in the query. It should be after the closing bracket that closes the WHERE clause:

    CONSTRUCT  { ?s ?p ?o } 
    WHERE 
    {
     GRAPH ?g { ?s ?p ?o } 
     ?s <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://lod.isi.edu/ontology/syllabus/Homework> .
     ?s <http://lod.isi.edu/ontology/syllabus/hasEventDate> ?date .
    }
    ORDER BY ASC(?date)
    

    Also note that several of the curly braces in your original query are, although not exactly wrong, superfluous.