On 28/02/2017 10:51, Andy Seaborne wrote: > What people on the list receive is: > > https://lists.apache.org/api/source.lua/8d17e5299cbbe585e7d45ff8ce156021162a9e16fe77818d8f7cff6c@%3Cusers.jena.apache.org%3E >
Ok. So the list is text based ;) Fine for next time. > > Looks like HTML to text conversion. > > On 27/02/17 23:58, George News wrote: >> But the funny thing is that in both cases (with and without UNION) >> the code is the same. I know I close the transaction, but I don’t >> understand why in one case it works and in the other no. > > Probably you close the transaction before the resultset is all > consumed. > > Some work is done on execSelect so internally, the use of the > dataset can finish early. > > UNION has two branches calculated separately. I was supposing that. At the end I have included the dataset in my own ResultSetClosable class, so this way I can call dataset.end() on ResultSet close. It seems that it works this way, it is not the more elegant one, but it works. In this sense, one option for including in Jena could be to enable the option to create a QueryExecution with transaction included, in order to avoid some issues like that. > It depends on the real query. > > Andy > >> >> I would like to ask another thing: In a @requestscope is the >> reading transaction automatically close when finished the REST >> processing or is it mandatory to call dataset.end()? >> >> I don't knw why example code is unreadable. I have perfectly seen >> the code. >> >> Thanks a lot Jorge >> >> Sent from jlanza_lumia820 >> >> From: Andy Seaborne Sent: martes, 28 de febrero de 2017 0:12 To: >> [email protected] Subject: Re: SPARQL with UNION returning >> TDBTransactionException >> >> Unreadable. >> >> ------------------------- >> >> Probably you are passing the result stream out of the transaction. >> >> Reading a ResultSet requires reading the dataset and must be inside >> a transaction when the calls to hasNext/next are happening. >> >> Andy >> >> On 27/02/17 18:18, George News wrote: >>> |public SparqlResult executeSparql(String sparql) throws >>> SparqlExecutionException { String queryString = sparql; Query >>> query = QueryFactory.create(queryString); >>> dataset.begin(ReadWrite.READ); try { QueryExecution qExec = >>> QueryExecutionFactory.create(query, getModel()); // >>> System.out.println(qExec.getQuery().serialize()); // >>> System.out.println(qExec.getQuery().toString()); SparqlResult >>> result; if (query.isSelectType()) { result = new >>> SparqlResult(qExec.execSelect(), qExec); } else if >>> (query.isDescribeType()) { result = new >>> SparqlResult(qExec.execDescribe(), qExec); } else if >>> (query.isAskType()) { result = new SparqlResult(qExec.execAsk(), >>> qExec); } else if (query.isConstructType()) { result = new >>> SparqlResult(qExec.execConstruct(), qExec); } else { throw new >>> SparqlExecutionException("Unsupported query type: " + >>> query.getQueryType()); } return result; } finally { >>> dataset.end(); } private Map<String, Integer> >>> getSummaryStatistics() { String queryString = "THE_ONE"; >>> Map<String, Integer> statistics = new HashMap<>(); >>> GlobalTripleStore gts = new GlobalTripleStore(); try >>> (SparqlResult result = gts.executeSparql(queryString)) { >>> ResultSet resultSet = (ResultSet) result.getResult(); while >>> (resultSet.hasNext()) { QuerySolution sol = resultSet.next(); int >>> devices = sol.get("count_devices").asLiteral().getInt(); int >>> observations =sol.get("count_o").asLiteral().getInt(); >>> statistics.put("devices", resources); statistics.put("o", >>> observations); } } catch (SparqlExecutionException e) { // TODO >>> Auto-generated catch block e.printStackTrace(); } return >>> statistics; } | >> >> >
