Strange route for fastest car

Hi,

I found a very interesting routing result for fastest car that I cannot explain. I originally thought it was due to data quality but it seems not.

The route is here:
https://graphhopper.com/maps/?point=-24.746442%2C119.59794&point=-24.717831%2C119.605837&locale=en-US&vehicle=car&weighting=fastest&elevation=true&use_miles=false&layer=Omniscale

And this is OSM way that covers both start and end:

Can anyone please explain how this could happen?

Cheers,
Tuan.

if i understand correctly, the reason is that the following two ford nodes cut off the way and the part of the way between them is removed from the graph.

1 Like

Yes this is correct, also see this issue.

Thanks jie31best for pointing out the issue. I tried generating graph with block_fords=false and I was able to find the shortest path.

And thanks boldtrn for providing the context on this.

1 Like

We are working on a more generic solution where we store properties like street type or ford etc in the graph, which will 1) allow a change in the routing calculation per request (if CH is disabled) and 2) gives this information to the client via more PathDetails so if you route through fords you can notify your users :slight_smile:

1 Like

This refactoring will be really useful Peter. Thanks for the information. Looking forward to its release :slight_smile:

1 Like