|
Welcome Guest | Login | |
|
Index
| Recent Threads
| Who's Online
| User List
| Search
| Help
| 
|
|
Forum has been closed down This forum has been closed down due to extensive spamming activities. Please use the mailing list instead. |
 |
openRDF.org Forum » Sesame & Rio: Help » Thread: Two problems with openrdf-workbench (2.0beta4) |
|
Total posts in this thread: 5 |
[Add To My Favorites] [Watch this Thread] |
| Author |
|
|
Jul 1, 2007 10:31:22 PM
ScottMorrison Member
|
First, workbench doesn't seem to remember my choices of server and repository for very long; I'm continually being asked to reselect these, even though I click the checkbox to remember my choices. Second, workbench takes an extremely long time to load after selecting a large repository. My guess (?!) is that it is traversing the entire list of triples to build the list of contexts to display, but I have no real evidence for that. Combined, I can't perform queries on a large repository; by the time workbench has gotten around to showing me the list of contexts, and I'm able to click on the "Query" link on the right, it's decided to forget my choice of server and repository. Thus clicking on "Query" just starts the cycle again! Thanks, Scott Morrison http://tqft.net/ |
||
|
|
Jul 3, 2007 12:24:13 PM
arjohn OpenRDF project lead The Netherlands Joined: Jan 23, 2004 Posts: 1289 Status: Offline |
We just verified this, it's a bug. I created an issue for this in the issue tracker: [SES-419].
Most likely cause is the retrievel of the list of contexts. Can you indicate how long "extremely long" is in minutes/hours? Also, it would help if we know what type of storage your server uses (e.g. native, memory). ---------------------------------------- Arjohn Kampman, OpenRDF project lead, Aduna |
|||||
|
|
Jul 3, 2007 3:12:37 PM
Herko_ter_Horst Sesame developer
|
Both issues have been resolved. With regards to Workbench not remembering the server: this was due to the fact that the cookie used to store the server information was not set to persist, so the info was lost when the browser was closed. In addition, it turned out that there was a bug reading the cookie information. This has now been fixed. I've also removed the display of the list of all contexts from the repository overview page. This obviously doesn't scale at all. In addition, the information isn't very useful in any case. |
||
|
|
Aug 31, 2007 12:56:30 AM
ScottMorrison Member
|
The replies above explain that listing all contexts doesn't scale well. However, this is something I forsee wanting to do! Is it true that simply by adding a "context-first" index, for example "cspo", I can get around this? That is, does a "cspo" index allow the NativeStore implementation to list contexts in time proportional to number of contexts, rather than number of statements? Thanks, Scott Morrison |
||
|
|
Aug 31, 2007 8:51:51 AM
arjohn OpenRDF project lead The Netherlands Joined: Jan 23, 2004 Posts: 1289 Status: Offline |
I'm afraid that simply adding a cspo index currently won't reduce the time required to get all context resources. Such an index does have the potential to increase performance, but this optimization has not yet been implemented. ---------------------------------------- Arjohn Kampman, OpenRDF project lead, Aduna |
|||
|
| [Show Printable Version of Thread] |
