GraphHopper.com | Forum | GitHub | Maps | Blog

How to find the offending data with "Speed cannot be 0 for unblocked edge"


#1

There is a fairly big area in southern Germany that we are unable to get isochrones calculated in.
Each of these gives us the error below:

curl "http://localhost:8989/isochrone?point=48.89,9."{15..37}
curl "http://localhost:8989/isochrone?point=48."{86..94}",9.20"

Speed cannot be 0 for unblocked edge, use access properties to mark edge blocked! Should only occur for shortest path calculation. See #242.

I did investigate the area in OSM’s live data for anything tagged speed, maxspeed or minspeed =0 using Overpass Turbo but could not find anything. How could we proceed looking for the culprit? Which other values, tags or tag combinations would lead to a speed 0 in graphhopper?

Using the latest version of https://github.com/graphhopper/graphhopper/ (isochrone_open_source branch).
Tested with europe_germany.pbf of 2018-02-22 as well as 2018-03-06.


#2

Ah, thanks for posting here :slight_smile:

I think this is some bug. Can you try adding

if (Double.isInfinite(tmpWeight))
continue;

directly after

double tmpWeight = weighting.calcWeight(iter, reverseFlow, currEdge.edge) + currEdge.weight;

in Reachability.java?


#3

Thanks!

In isochrone/src/main/java/com/graphhopper/isochrone/algorithm/Isochrone.java you mean, right? That’s the only occurence of “double tmpWeight” I could find.

I removed my tools/target directory and ran ./graphhopper.sh web europe_germanpbf again. Not sure that this is all what’s needed for a clean rebuild?

It said

building graphhopper jar: tools/target/graphhopper-tools-0.11-SNAPSHOT-jar-with-dependencies.jar

I get the same results as before.


#4

Ah, yes. Sorry we renamed it in the branch.

I removed my tools/target directory and ran ./graphhopper.sh web europe_germanpbf again. Not sure that this is all what’s needed for a clean rebuild?

Can you try with an IDE or:

./graphhopper.sh clean
./graphhopper.sh web europe_german.pbf

?


#5

did it!

It works and the isochrones look reasonable.

Thank you so much!