I am working on a project (java using NetBeans) which at some point needs to process map data. I have got Esri shape files for whole UK and need to create one dimensional index using geoHash as efficiency is a big factor in my project. Since I need to search distances I am going to use Morton ordering. However the problem is that geohash-ing for close points on each side of equator can results in hash indices with no common prefix which in turn messes up my proximity searches in my binary search tree.
Has any body have neat way of doing this? I have a way which is adding an offset for all points in Uk but this solution is not nicely scalable. I know that MongoDb is using geoHash and still doing proximity searches, so should be away already implemented.