linkinghere module: string_output_string is longer than 10Mb

[seebi]

some sparql query in the linking here module is not optimal for large datasets:
SPARQL Error: [unixODBC][OpenLink][Virtuoso iODBC Driver][Virtuoso Server]HT057: The STRING session in string_output_string is longer than 10Mb. Either use substring to access it in parts or place less data in it. () in query: define output:format "JSON" SELECT DISTINCT ?subject ?uri FROM http://leipzig-data.de/Data/Adressen/ WHERE { ?subject ?uri http://leipzig-data.de/Data/Model/Adresse . }

as seen here: http://leipzig-data.de/Data/resource/properties?r=http%3A%2F%2Fleipzig-data.de%2FData%2FModel%2FAdresse

[seebi]

maybe we limit the query to 6? (more is not shown here)

[white-gecko]

If I understand you right you suggest the following patch:

--- a/extensions/resourcemodules/LinkinghereModule.php
+++ b/extensions/resourcemodules/LinkinghereModule.php
@@ -33,7 +33,7 @@ class LinkinghereModule extends OntoWiki_Module
                 'WHERE {
                     ?subject ?uri <' . (string)$this->_owApp->selectedResource . '> .
                 }'
-            );
+            )->setLimit(6);

         $result           = $this->_owApp->selectedModel->sparqlQuery($query, array('result_format' => 'extended'));
         $_predicatesResult = array();

I can't test it locally because the exports ends with the following exception:

SPARQL Error: [unixODBC][OpenLink][Virtuoso iODBC Driver][Virtuoso Server]SR353: Sorted TOP clause specifies more then 11000 rows to sort. Only 10000 are allowed. Either decrease the offset and/or row count or use a scrollable cursor () in query: …

It seams that we have more scaling problems :-(

[seebi]

@hgg: please upgrade your installation and report success or fail ...

[cfrancois7]

Hello.

I got the same error than you with the last version.

SELECT ?s ?p
FROM <http://www.ecoinvent.org/2013/ecospold#>
WHERE{?s ?p <http://www.ecoinvent.org/2013/ecospold#flowType> . }  -> 

/usr/local/var/www/htdocs/vendor/aksw/erfurt/library/Erfurt/Store/Adapter/Virtuoso.php@1058
2017-03-28T08:52:18+02:00 EMERG (0): Erfurt_Store_Adapter_Exception: 
SPARQL Error: [OpenLink][Virtuoso iODBC Driver][Virtuoso Server]HT057:
The STRING session in string_output_string is longer than 10Mb.
Either use substring to access it in parts or place less data in it.
() on querying graph <NULL> with query: define output:format "JSON"

While the limit is set in the linkingheremodule:

$query->setSelectClause('SELECT DISTINCT ?subject ?uri')
            ->setWherePart(
                'WHERE {
                    ?subject ?uri <' . (string)$this->_owApp->selectedResource . '> .
                }'
            )->setLimit(6);

[cfrancois7]

I found something else related to this error:

[Tue Mar 28 09:19:29.118377 2017] [:error] [pid 7715] [client ::1:54873] PHP Fatal error: Maximum execution time of 300 seconds exceeded in /usr/local/var/www/htdocs/vendor/aksw/erfurt/library/Erfurt/Rdf/MemoryModel.php on line 382, referer: http://localhost/index.php/resource/properties/?r=http%3A%2F%2Fwww.ecoinvent.org%2F2013%2Fecospold%23flowType%2FreferenceProduct

This node referenceProduct is a very linked instance, a kind of super-identifier of thousands other nodes.

I got the same error with another similar type of instance :

Fatal error: Maximum execution time of 300 seconds exceeded in /usr/local/var/www/htdocs/vendor/aksw/erfurt/library/Erfurt/Rdf/MemoryModel.php on line 382

The error appears when I load an instance that is linked to a such "very linked" instance.
How is possible to limit the loading of data?