question about use LUBM

[彼岸 <ki...@qq.com>]

Hello ,everyone:
       Does anyone familiar with   LUBM ( http://swat.cse.lehigh.edu/projects/lubm/)  ?
       Today, I just downloaded the data generator  UBA1.7 to get the data. I haven't use any parameter except "-onto" .After the program run is completed  ,I got 14 output data files ,the name as  University0_0.owl ,University0_1.owl  to  University0_14.owl . 
       At present , I just use the University0_0.owl  as the input test data for the 14 sparql queries(These queries is provided by LUBM) . My question is how to use the rest 13 data files . Dose every file for one sparql query? such as  University0_1.owl for Query2 ,University0_2 for Query 3 and so on.
       
       Best regards to all !

Re: question about use LUBM

[Andy Seaborne <an...@apache.org>]

On 30/04/14 05:25, 彼岸 wrote:
> Hello ,everyone: Does anyone familiar with   LUBM (
> http://swat.cse.lehigh.edu/projects/lubm/)  ? Today, I just
> downloaded the data generator  UBA1.7 to get the data. I haven't use
> any parameter except "-onto" .After the program run is completed  ,I
> got 14 output data files ,the name as  University0_0.owl
> ,University0_1.owl  to  University0_14.owl . At present , I just use
> the University0_0.owl  as the input test data for the 14 sparql
> queries(These queries is provided by LUBM) . My question is how to
> use the rest 13 data files . Dose every file for one sparql query?
> such as  University0_1.owl for Query2 ,University0_2 for Query 3 and
> so on.
>
> Best regards to all !
>

Load all the files and query the combined graph and run the benchmark 
queries against that.

LUBM does not have any significant cross links between datasets so it's 
a bit limited in it's usefulness.  It loads about 2x faster than other 
data as well (everyone finds this - it's not specific to Jena) due to 
the link/node ration so also treat load benchmarks using LUBM with 
caution as well.

What are you trying to measure?

	Andy