| 
     
      
      
      From: Michael S. <ms...@me...> - 2015-05-15 17:20:45
      
     
   | 
Dear Lee, thanks for reporting the two issues! We have identified the problems causing the behavior (see my comment in the parent task http://jira.blazegraph.com/browse/BLZG-1141 <http://jira.blazegraph.com/browse/BLZG-1141>). A correctness fix for both issues is available in branch BLZG48and50 <https://github.com/SYSTAP/bigdata/tree/BLZG48and50>, however the fix might induce some performance regressions. We want to look at performance in more depth before merging into master (probably as part of a more exhaustive refactoring), so it might take a while until this will appear in master. Best, Michael > On 13 May 2015, at 16:06, Bryan Thompson <br...@sy...> wrote: > > Lee, > > Many thanks. I will assign them to Michael (Cc) to take an initial look. > > Thanks, > Bryan > ---- > Bryan Thompson > Chief Scientist & Founder > SYSTAP, LLC > 4501 Tower Road > Greensboro, NC 27410 > br...@sy... > http://blazegraph.com > http://blog.bigdata.com > http://mapgraph.io > > Blazegraph™ is our ultra high-performance graph database that supports > both RDF/SPARQL and Tinkerpop/Blueprints APIs. MapGraph™ is our > disruptive new technology to use GPUs to accelerate data-parallel > graph analytics. > > CONFIDENTIALITY NOTICE: This email and its contents and attachments > are for the sole use of the intended recipient(s) and are confidential > or proprietary to SYSTAP. Any unauthorized review, use, disclosure, > dissemination or copying of this email or its contents or attachments > is prohibited. If you have received this communication in error, > please notify the sender by reply email and permanently delete all > copies of the email and its contents and attachments. > > > On Wed, May 13, 2015 at 10:03 AM, Lee Kitching <le...@sw...> wrote: >> Hi Bryan, >> >> I've raised these at http://jira.blazegraph.com/browse/BLZG-48 and >> http://jira.blazegraph.com/browse/BLZG-50. >> >> Thanks >> >> On Wed, May 13, 2015 at 12:41 PM, Bryan Thompson <br...@sy...> wrote: >>> >>> Lee, >>> >>> These definitely look like bugs. Can you please file tickets at >>> http://jira.blazegraph.com and include sample data and expected >>> solutions for each? We will try to get the fixes into the 1.5.2 >>> release. >>> >>> Thanks, >>> Bryan >>> ---- >>> Bryan Thompson >>> Chief Scientist & Founder >>> SYSTAP, LLC >>> 4501 Tower Road >>> Greensboro, NC 27410 >>> br...@sy... >>> http://blazegraph.com >>> http://blog.bigdata.com >>> http://mapgraph.io >>> >>> Blazegraph™ is our ultra high-performance graph database that supports >>> both RDF/SPARQL and Tinkerpop/Blueprints APIs. MapGraph™ is our >>> disruptive new technology to use GPUs to accelerate data-parallel >>> graph analytics. >>> >>> CONFIDENTIALITY NOTICE: This email and its contents and attachments >>> are for the sole use of the intended recipient(s) and are confidential >>> or proprietary to SYSTAP. Any unauthorized review, use, disclosure, >>> dissemination or copying of this email or its contents or attachments >>> is prohibited. If you have received this communication in error, >>> please notify the sender by reply email and permanently delete all >>> copies of the email and its contents and attachments. >>> >>> >>> On Wed, May 13, 2015 at 6:51 AM, Lee Kitching <le...@sw...> wrote: >>>> Hi, >>>> >>>> We are trying to migrate our native Sesame database to Blazegraph, >>>> however >>>> we are encountering a problem with the following query which returns no >>>> results in Blazegraph: >>>> >>>> SELECT DISTINCT ?uri WHERE { >>>> { >>>> SELECT DISTINCT (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> VALUES ?dg { <http://graph1> } >>>> GRAPH ?dg { ?uri_d a ?type . } >>>> } >>>> } >>>> MINUS >>>> { >>>> SELECT DISTINCT (?uri_l as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> VALUES ?lg { <http://graph2> } >>>> GRAPH ?lg { ?uri_l a ?type . } >>>> } >>>> } >>>> } >>>> >>>> This returns a single result in Sesame since the first subquery returns >>>> one >>>> result and the subquery in the MINUS clause returns none. Executing the >>>> first subquery >>>> >>>> SELECT DISTINCT (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> VALUES ?dg { <http://graph1> } >>>> GRAPH ?dg { ?uri_d a ?type . } >>>> } >>>> >>>> returns 0 results in Blazegraph, however moving the single graph value >>>> into >>>> the GRAPH clause: >>>> >>>> SELECT DISTINCT (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> GRAPH <http://graph1> { ?uri_d a ?type . } >>>> } >>>> >>>> returns the expected result. Combining the two VALUES clauses into one: >>>> >>>> SELECT DISTINCT (?uri_d as ?uri) WHERE { >>>> VALUES (?type ?dg) { (<http://type1> <http://graph1>) (<http://type2> >>>> <http://graph1>) (<http://type3> <http://graph1>) } >>>> GRAPH ?dg { ?uri_d a ?type . } >>>> } >>>> >>>> also returns the result. Does Blazegraph support multiple VALUES clauses >>>> in >>>> the same query? I would expect the three queries above to be equivalent >>>> (in >>>> the case of a single graph). >>>> >>>> Another issue is that Blazegraph appears to be inconsistent when >>>> projecting >>>> variables from inner subqueries. The query: >>>> >>>> SELECT DISTINCT (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> GRAPH <http://graph1> { ?uri_d a ?type . } >>>> } >>>> >>>> returns the expected result, however wrapping this in another >>>> projection: >>>> >>>> SELECT ?uri WHERE { >>>> SELECT (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> GRAPH <http://graph1> { ?uri_d a ?type . } >>>> } >>>> } >>>> >>>> causes no results to be returned. Surprisingly, projecting an unused >>>> variable from the inner query: >>>> >>>> SELECT ?uri WHERE { >>>> SELECT ?type (?uri_d as ?uri) WHERE { >>>> VALUES ?type { <http://type1> <http://type2> <http://type3> } >>>> GRAPH <http://graph1> { ?uri_d a ?type . } >>>> } >>>> } >>>> >>>> causes the result to be returned again. >>>> >>>> These two issues appear to be preventing the original MINUS query from >>>> returning the expected results. Could you confirm these behaviours are >>>> bugs? >>>> I tried looking on the issue tracker but couldn't find anything directly >>>> related. >>>> >>>> Thanks >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> One dashboard for servers and applications across Physical-Virtual-Cloud >>>> Widest out-of-the-box monitoring support with 50+ applications >>>> Performance metrics, stats and reports that give you Actionable Insights >>>> Deep dive visibility with transaction tracing using APM Insight. >>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y >>>> _______________________________________________ >>>> Bigdata-developers mailing list >>>> Big...@li... >>>> https://lists.sourceforge.net/lists/listinfo/bigdata-developers >>>> >> >>  |