CODEWITHSUNDEEP
sparqljena
Luciane
Luciane

Reputation: 259

Best Practices on querying for multiple properties of a resource in the same SPARQL query

On my database, I have triples like:

DocumentUri -> dc.title -> title 
DocumentUri -> dc.language -> language 
DocumentUri -> dc.description -> description 
DocumentUri -> dc.creator -> AuthorUri

I'd like to be able to search for a documenttitle and then get all the properties from all the documents matching the title search.

I'm trying to do that with Jena and SPARQL. I made a query that receives a title to get the Uris from the documents that have the given title. That's the method, it gets the uris returned and store them in a list called webDocumentListInicial:

public void searchUriByTitle() {
        RDFNode documentUriNode;

        String queryString = "PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> " +
                "PREFIX dc: <http://purl.org/dc/elements/1.1/> SELECT ?document WHERE { " +
                "?document dc:title ?title." +
                "FILTER (?title = \"" + this.getTitle() + "\" ). }";

        Query query = QueryFactory.create(queryString);

        QueryExecution qe = QueryExecutionFactory.create(query, databaseModel);
        ResultSet results =  qe.execSelect();

        while( results.hasNext() ) {

           QuerySolution querySolution = results.next();
           documentUriNode = querySolution.get("document");

           WebDocument document = new WebDocument(documentUriNode.toString());
          this.webDocumentListInicial.add(document);

        }

        qe.close();  
    }

To get the document's creator I made another query, because in this case tha value from the triple is another resource. Here, I iterate the list of document URIs that was filled in the method above.

public void searchAuthorByTitle() {
    for(  WebDocument doc : this.webDocumentListInicial ) {
    RDFNode authorUriNode;

    String queryString = "PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> " +
            "PREFIX dc: <http://purl.org/dc/elements/1.1/> SELECT ?author WHERE { " +
            "?document dc:creator ?author." +
            "FILTER (?document = <" + doc.getUri() + "> ). }";

    Query query = QueryFactory.create(queryString);

    QueryExecution qe = QueryExecutionFactory.create(query, databaseModel);
    ResultSet results =  qe.execSelect();

    while( results.hasNext() ) {

       QuerySolution querySolution = results.next();
       authorUriNode = querySolution.get("author");

       WebAuthor author; 
       author = this.searchAuthorProperties(authorUriNode.toString(), new WebAuthor(authorUriNode.toString()) );

       doc.addAuthor(author);
    }
    qe.close();  
    }
}

And to get the other document properties, I do like in the example below, where I iterate the list that was filled in the first method I showed above.

public void searchDescription() {

        for( WebDocument doc : this.webDocumentListInicial ) {
            String description = "";

            Resource resource = ResourceFactory.createResource(doc.getUri());
            StmtIterator descriptionStmtIt = databaseModel.listStatements(resource, DC.description,(RDFNode) null);

            while( descriptionStmtIt.hasNext() ) {
                description = descriptionStmtIt.next().getObject().toString();
            }
            doc.setDescription(description);
        } 

    }

This way I'm handling with the data isn't very productive because I need a different query for each property I get.

Is it possible to make only one query to get the document URI and all the other document's properties at once? I tried that once, like this:

String queryString = "PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> " +
                "PREFIX dc: <http://purl.org/dc/elements/1.1/> SELECT ?document ?description " +
                "?language ?author  WHERE { " +
                "?document dc:title ?title." +
                "?document dc.language ?language" +
                "?document dc.description ?description" +
                "?document dc.creator ?author" +
                "FILTER (?title = \"" + this.getTitle() + "\" ). }";

But when I had more than one document matching the given title, it was difficult to know which properties returned belonged to each document.

Thank you!!

Upvotes: 2

Views: 1966

Answers (1)

Joshua Taylor
Joshua Taylor

Reputation: 85843

Building a better query

It sounds like you're doing a lot more work than you need to. If you have data like this:

@prefix : <http://stackoverflow.com/q/20436820/1281433/>

:doc1 :title "Title1" ; :author :author1 ; :date "date-1" .
:doc2 :title "Title2" ; :author :author2 ; :date "date-2" .
:doc3 :title "Title3" ; :author :author3 ; :date "date-3" .
:doc4 :title "Title4" ; :author :author4 ; :date "date-4" .
:doc5 :title "Title5" ; :author :author5 ; :date "date-5" .

And a list of titles, say "Title1" "Title4" "Title5" and you want retrieve the resource of the document with each title, along with the associated author and date, you can use a query like this:

prefix : <http://stackoverflow.com/q/20436820/1281433/>

select ?document ?author ?date where {
  values ?title { "Title1" "Title4" "Title5" }

  ?document :title ?title ;
            :author ?author ;
            :date ?date .
}

You'll get results like this in one ResultSet. There's no need to make multiple queries.

----------------------------------
| document | author   | date     |
==================================
| :doc1    | :author1 | "date-1" |
| :doc4    | :author4 | "date-4" |
| :doc5    | :author5 | "date-5" |
----------------------------------

Building a map of the results

Based on your comments, it sounds like you need to construct some other kind of associative structure from the ResultSet. Here's one way that you could construct a Map<RDFNode,Map<String,RDFNode>> that takes each document IRI to another map that takes each of the variable named to the associated value.

import java.io.ByteArrayInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;

import com.hp.hpl.jena.query.QueryExecutionFactory;
import com.hp.hpl.jena.query.QuerySolution;
import com.hp.hpl.jena.query.ResultSet;
import com.hp.hpl.jena.rdf.model.Model;
import com.hp.hpl.jena.rdf.model.ModelFactory;
import com.hp.hpl.jena.rdf.model.RDFNode;

public class HashedResultsExample {

    final static String DATA =
            "@prefix : <http://stackoverflow.com/q/20436820/1281433/>\n" +
            "\n" +
            ":doc1 :title 'Title1' ; :author :author1 ; :date 'date-1' .\n" +
            ":doc2 :title 'Title2' ; :author :author2 ; :date 'date-2' .\n" +
            ":doc3 :title 'Title3' ; :author :author3 ; :date 'date-3' .\n" +
            ":doc4 :title 'Title4' ; :author :author4 ; :date 'date-4' .\n" +
            ":doc5 :title 'Title5' ; :author :author5 ; :date 'date-5' .\n" ;

    final static String QUERY = 
            "prefix : <http://stackoverflow.com/q/20436820/1281433/>\n" +
            "select ?document ?author ?date where {\n" +
            "  values ?title { \"Title1\" \"Title4\" \"Title5\" }\n" +
            "  ?document :title ?title ; :author ?author ; :date ?date .\n" +
            "}" ;

    public static void main(String[] args) throws IOException {
        final Model model = ModelFactory.createDefaultModel();
        try ( final InputStream in = new ByteArrayInputStream( DATA.getBytes() )) {
            model.read( in, null, "TTL" );
        }

        final ResultSet rs = QueryExecutionFactory.create( QUERY, model ).execSelect();
        final Map<RDFNode,Map<String,RDFNode>> map = new HashMap<>();

        while ( rs.hasNext() ) {
            final QuerySolution qs = rs.next();
            final Map<String,RDFNode> rowMap = new HashMap<>();
            for ( final Iterator<String> varNames = qs.varNames(); varNames.hasNext(); ) {
                final String varName = varNames.next();
                rowMap.put( varName, qs.get( varName ));
            }
            map.put( qs.get( "document" ), rowMap );
        }

        System.out.println( map );
    }
}

The output (since the map is printed at the end) with some newlines for readability is:

{http://stackoverflow.com/q/20436820/1281433/doc4=
 {author=http://stackoverflow.com/q/20436820/1281433/author4,
  document=http://stackoverflow.com/q/20436820/1281433/doc4,
  date=date-4},
 http://stackoverflow.com/q/20436820/1281433/doc1=
 {author=http://stackoverflow.com/q/20436820/1281433/author1,
  document=http://stackoverflow.com/q/20436820/1281433/doc1,
  date=date-1},
 http://stackoverflow.com/q/20436820/1281433/doc5=
 {author=http://stackoverflow.com/q/20436820/1281433/author5,
  document=http://stackoverflow.com/q/20436820/1281433/doc5,
  date=date-5}}

Upvotes: 4

Related Questions