Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [geomesa-users] new index -> new implementation for knn-query?

Dear Marcel,
Nice graphics! Your second figure is indeed correct: the KNN algorithm will search all geohashes which could possibly contain points closer than those already found. However, the first figure is not accurate, but the fault is mine in choosing a misleading name: the GeoHashSpiral is not necessarily a spiral made of GeoHashes as your figure indicated; rather it is a iterator which yields GeoHashes ordered by their *minimum* geodetic distance from the reference point. Because of this, the order in which GeoHashes are searched will "jitter" around, with a striking (yet correct) pattern occurring near the poles in some cases. Is my explanation clear?

You are correct that the KNN process must be re-implemented in order to utilize the new z3 index. I ( as the original author ) need to sit down with some other contributors to see how and when this (quite fun) work might be done.

Keep the questions coming,
Mike Ronquest



On 09/24/2015 04:59 AM, Marcel wrote:
Hello,
at first I want to thank you all for the last explanations of my questions. The support is really fast and motivated :)
This leads me to the next question regarding the new index:
When geohashes were not used anymore, a new implemenation for the knn query should be necessary, because it uses a geohashspiral. In the attachment you find my understanding for the current execution of a knn query. First picture shows the geohashspiral and second picture shows the test whether all points found are really the closest ones. For this all skin-colored geohashes were processed, because they could contain some points that are closer to my origin. The small circle contains the result for k=10. I could imagine that these two pictures only show a simplified version of the algorithm, but it would be great if you confirm my understanding is correct. In my masterthesis a want to explain these method a bit more in detail. So when a new version comes up, how will this new implemenation look like? Maybe by starting with a small circle which is growing continously?

Thanks a lot,
Marcel Jacob.


_______________________________________________
geomesa-users mailing list
geomesa-users@xxxxxxxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
http://www.locationtech.org/mailman/listinfo/geomesa-users



Back to the top