If I have to reload the entire PBF file can that be done whilst the server is running, or must I stop it, clear out the data (?) and reload it ?
It is required to drop the graph folder and do an entire re-import yes. Still multiple graphhopper instances can run on the same server (using a different port, or with multiple GraphHopper.java instances even in one JVM but this requires change in the source code, see https://github.com/graphhopper/graphhopper/issues/211). Furthermore you can even remove the graph folder while the service is running, if you are using graph.dataaccess: RAM_STORE
Hi Peter,
Many thanks for your response.
Ok, that’s not quite what I want as I just sought to grab the changes deltas. These are only available in XML format as far as I can tell. So it’s not quite joining multiple PBF’s (unless I am missing something).
Being able to partially update the files from either changefiles, or perhaps simpler, an updated pbf for a given region would be an amazing feature to have. When using graphhopper at the continental scale it becomes quite costly to generate a new graph.