Spatial RDD

Custom Spark RDD to partition geospatial data, based on spatial proximity, for faster orthogonal range queries.

What is this Fork about?

This fork modifies the Spatial RDD to partition dataset using KD Tree & Epsilon approximation based on Parallel Algorithms for Constructing Range and Nearest-Neighbor Searching Data Structures.

Note

1. Here we have chosen to implement KD tree based on 2D points
2. We are doing primary partitioning & secondary indexing using KD Tree.