You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@clerezza.apache.org by "Oliver Strässer (JIRA)" <ji...@apache.org> on 2010/04/21 11:08:50 UTC
[jira] Created: (CLEREZZA-194) Replace the integrated SPARQL engine
Replace the integrated SPARQL engine
------------------------------------
Key: CLEREZZA-194
URL: https://issues.apache.org/jira/browse/CLEREZZA-194
Project: Clerezza
Issue Type: Improvement
Reporter: Oliver Strässer
Priority: Critical
The Scalability of the integrated SPARQL engine isn't very good. Some test with big graphs demonstrate this problem.
A Query over a big Graph (120 MB) needs more than 45 seconds with SPARQL. The same "query" build with graph.filter(triple) calls, needs only 1-2 msec - quite faster.
research with Manuel Innerhofer shows, that the problem is the hasnext() method of the SPARQL resultset.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Closed: (CLEREZZA-194) Replace the integrated SPARQL engine
Posted by "Reto Bachmann-Gmür (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CLEREZZA-194?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Reto Bachmann-Gmür closed CLEREZZA-194.
---------------------------------------
Assignee: Reto Bachmann-Gmür
Resolution: Won't Fix
While a Fastlane might improve perfoamce, no sparql engine is known that is likely to be faster than arq
> Replace the integrated SPARQL engine
> ------------------------------------
>
> Key: CLEREZZA-194
> URL: https://issues.apache.org/jira/browse/CLEREZZA-194
> Project: Clerezza
> Issue Type: Improvement
> Reporter: Oliver Strässer
> Assignee: Reto Bachmann-Gmür
> Priority: Critical
>
> The Scalability of the integrated SPARQL engine isn't very good. Some test with big graphs demonstrate this problem.
> A Query over a big Graph (120 MB) needs more than 45 seconds with SPARQL. The same "query" build with graph.filter(triple) calls, needs only 1-2 msec - quite faster.
> research with Manuel Innerhofer shows, that the problem is the hasnext() method of the SPARQL resultset.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Issue Comment Edited: (CLEREZZA-194) Replace the integrated
SPARQL engine
Posted by "Reto Bachmann-Gmür (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CLEREZZA-194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860886#action_12860886 ]
Reto Bachmann-Gmür edited comment on CLEREZZA-194 at 4/26/10 6:41 AM:
----------------------------------------------------------------------
It would be inetersting to have a comparison of the speed for a sparql query directly against TDB and the same query using clerezza. As the possible perfoamce improvement would be to implement the envisage fastlane to a sparql endpoint provided by the storage backend.
was (Author: reto):
It would be inetersting to have a comparison of the speed for a sparql query directly against TDB and the same query using clerezza.
> Replace the integrated SPARQL engine
> ------------------------------------
>
> Key: CLEREZZA-194
> URL: https://issues.apache.org/jira/browse/CLEREZZA-194
> Project: Clerezza
> Issue Type: Improvement
> Reporter: Oliver Strässer
> Priority: Critical
>
> The Scalability of the integrated SPARQL engine isn't very good. Some test with big graphs demonstrate this problem.
> A Query over a big Graph (120 MB) needs more than 45 seconds with SPARQL. The same "query" build with graph.filter(triple) calls, needs only 1-2 msec - quite faster.
> research with Manuel Innerhofer shows, that the problem is the hasnext() method of the SPARQL resultset.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CLEREZZA-194) Replace the integrated SPARQL
engine
Posted by "Reto Bachmann-Gmür (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CLEREZZA-194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860886#action_12860886 ]
Reto Bachmann-Gmür commented on CLEREZZA-194:
---------------------------------------------
It would be inetersting to have a comparison of the speed for a sparql query directly against TDB and the same query using clerezza.
> Replace the integrated SPARQL engine
> ------------------------------------
>
> Key: CLEREZZA-194
> URL: https://issues.apache.org/jira/browse/CLEREZZA-194
> Project: Clerezza
> Issue Type: Improvement
> Reporter: Oliver Strässer
> Priority: Critical
>
> The Scalability of the integrated SPARQL engine isn't very good. Some test with big graphs demonstrate this problem.
> A Query over a big Graph (120 MB) needs more than 45 seconds with SPARQL. The same "query" build with graph.filter(triple) calls, needs only 1-2 msec - quite faster.
> research with Manuel Innerhofer shows, that the problem is the hasnext() method of the SPARQL resultset.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (CLEREZZA-194) Replace the integrated SPARQL
engine
Posted by "Manuel Innerhofer (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CLEREZZA-194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860888#action_12860888 ]
Manuel Innerhofer commented on CLEREZZA-194:
--------------------------------------------
I tried that. There was no big speed improvement observable.
> Replace the integrated SPARQL engine
> ------------------------------------
>
> Key: CLEREZZA-194
> URL: https://issues.apache.org/jira/browse/CLEREZZA-194
> Project: Clerezza
> Issue Type: Improvement
> Reporter: Oliver Strässer
> Priority: Critical
>
> The Scalability of the integrated SPARQL engine isn't very good. Some test with big graphs demonstrate this problem.
> A Query over a big Graph (120 MB) needs more than 45 seconds with SPARQL. The same "query" build with graph.filter(triple) calls, needs only 1-2 msec - quite faster.
> research with Manuel Innerhofer shows, that the problem is the hasnext() method of the SPARQL resultset.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.