blazegraph / database

Blazegraph High Performance Graph Database
GNU General Public License v2.0
898 stars 173 forks source link

Inference module and questions #185

Open kkutt opened 3 years ago

kkutt commented 3 years ago

I am writing to you, because I have a problem with inference module. Because nowadays I am working to custom solution where I use the update query 'CREATE ENTAILMENTS' and special INSERT query where manually create the triple in database.

query1 = '''INSERT

{ ?datacarrier tms:overrunSurvey ?observation . ?observation tms:overrunType tms:Max . } WHERE

{ ?datacarrier rdf:type tms:dataCarrier ; tms:hasLimitations [ tms:limitVariable ?variable; schema:maxValue ?max]. ?observation rdf:type sosa:Observation ; sosa:observedProperty ?variable ; sosa:hasSimpleResult ?value . FILTER( ?value > ?max ) } '''

and the next:

query2 = "CREATE ENTAILMENTS"

And my problem is the lock in database during the query2 execution which blocking insert a new observation sending by API from smartphone.

Can you help me the solution this problem? I wanna to run the INSERT during CREATE ENTAILMENTS. Is there a possibility to isolate these two queries?

I have a few requests to you, can you look my properties file (in the attachment), how can I tunning my properties file the better optimization database environment? I read your website (https://github.com/blazegraph/database/wiki/PerformanceOptimization) and I do not have anymore ideas. Where in the blazegraph files are saved information about the last database operations?

And last my problems is to reduce the size of jnl file. How can I do to this? Whether in the jnl file are saved the logs? Because I have only 37 millions triples and the size is 243 GB and this is too much for me.

Thanks in advance to you for your help in this matter. TMS_RWStore.properties.txt