Hmmh, good question. It does not happen for world wide and default encoders. It could be that this happened in earlier import methods with bugs. Should we throw an exception if this happens to find it out or should we remove this counter?
Regards, Peter On 22.07.2015 01:11, John Zhao wrote: > Hi, > > I look there is logic in PrepareRoutingSubnetworks.removeZeroDegreeNodes() > > I have tried a small area, like San Francisco. > And this is zero in the log. > 14:51:55.340 [main] INFO c.g.r.util.PrepareRoutingSubnetworks - > optimize to remove subnetworks (3500), zero-degree-nodes (0), > unvisited-dead-end-nodes(730), maxEdges/node (7) > > I remember, generally, a pillar node should have degree 2 at least. > (despite the dead end with degree 1.) > How it could be zero degree nodes in the graph? > > Can you give me an example? > > *Best Regards,* > *ZhiQiang ZHAO* > > > _______________________________________________ > GraphHopper mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/graphhopper
_______________________________________________ GraphHopper mailing list [email protected] https://lists.openstreetmap.org/listinfo/graphhopper
