Hi Siarhei, OrientDB privileges the traversal rather than insertion speed, because in 90% of the cases users do a massive import once and then use the database mostly on lookups.
However we're working to the new implementation of 1-N relationships mush faster. This will impact edge creation speed a lot. It should be available on 2.0-SNAPSHOT in few days. Lvc@ On 23 December 2013 21:32, Siarhei Liashkevich <[email protected]> wrote: > Hi All, > > I am trying to import really big data into OrientDB 1.6.2. Result graph > will include about 11M of vertexes and about 40M of edges with a lot of > many-to-many relationships. Currently my data stored in the CSV files, > linked via 'common' ids (to link vertexes together). To populate OrientDB > with data I am using self-written Java parser/convertor > (pure com.orientechnologies.orient.core.*, no blueprints). > Everything goes smoothly with vertexes, I can import them in a moderate > time as well as create 'common' index, so lookup (by 'common' id) is no a > problem. Edge creation went too slow. It took from 1 to 20 milliseconds to > create one edge (and looks like the time is growing), so the converter > can't finish the job in a moderate time. I read some post from this mail > thread, so I did an attempt to .save() not edges but nodes (all nodes at > the end, I don't have anything like root node), set OIntentMassiveInsert or > not to set it - no luck. > > So, the question is - does anybody knows the way to import such > configuration into OrientDB via Java converter ? > > -- > > --- > You received this message because you are subscribed to the Google Groups > "OrientDB" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/groups/opt_out. > -- --- You received this message because you are subscribed to the Google Groups "OrientDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
