search

harsha2010 / magellan

Geo Spatial Data Analytics on Spark
Apache License 2.0
532 stars 149 forks source link

Query for proximity search in a radius to find n-nearest #50

Open danfran opened 8 years ago

danfran commented 8 years ago

Is it possible with the current API (I didn't find a way yet), given a Lat/Lon (WGS84) point "A" and a set of similar coordinates, find for a specific radius (e.g.: 1 Mile) the first 100 nearest points around "A"?

In case it helps, in Mongo would be similar to: db.myData.find({ "location.coordinates" : { $near : { $geometry: { type: "Point", coordinates: [ 52.4702062,-0.2836129 ] }, $minDistance: 0, $maxDistance: 1600 } }, limit:100 })

harsha2010 commented 8 years ago

@danfran , we haven't added the distance measure to the API yet. It is in one of the branches and we will cut a new release within the next 3 weeks which should have this feature.

pvanickova commented 7 years ago

Did this feature make it to release harsha2010:magellan:1.0.4-s_2.11 ?

jak2030 commented 7 years ago

I'm also curious whether this is an undocumented or upcoming feature. Thank you!

harsha2010 commented 7 years ago

@jak2030 @pvanickova i haven't added this yet..nearest neighbor search is targeted for 1.0.6 (as it needs indexing to do this efficiently) I am working on nearest neighbor search this week, I expect a first cut PR to be ready in a week What is your use case? Are you willing to share some data to test and benchmark performance on? How many points do you need to compute kNN on, for example? What kind of resource constraints? Better yet, your use case description would be great as it will help in designing the algorithm