versions

1.0.0
  Initial release of SpatialHadoop with the following main features
   - Spatial data types: Point, Rectangle, Circle, Polygon and TigerShape
   - Spatial indexes: Grid and R-tree global indexes. R-tree local index.
   - Spatial MapReduce components:
     = SpatialRecordReader: Reads all types of blocks in spatial files (Heap, R-tree indexed)
     = BinaryRecordReader: Reads all types of blocks for binary operations
     = SpatialInputFormat: Base type for all types of input formats for spatial data. Uses BlockFilter to filter out unneeded blocks.
     = BinaryInputFormat: Base type of binary input formats. Used the binary version of BlockFilter to filter out unneeded blocks.
   - Spatial Operations:
     = RangeQuery: Performs a range query operation on both heap and spatially indexed files
     = KNN: Performs a KNN operation on both heap and spatially indexed files
     = DistributedJoin: Performs a spatial join on globally indexed files. Local index is not used yet.
     = SJMR: Performas a spatial join on heap files or globally indexed files.
     = FileMBR: Finds the MBR of a file
     = Randomizer: Randomizes an input file, i.e., shuffles the lines
     = Sampler: Reads an unbiased random sample from a file

1.0.1
  Minor bug fixes to ensure the result is the same whatever index structure or algorithm you use for the same dataset
  - Improved the RandomSpatialGenerator to avoid generating rectangles of zero width and/or height
  - Handle special cases with the RandomSpatialGenerator to correctly read lines at the boundaries of the block
  - Improve KNN to remove duplicates in the reduce phase
  - Improve writing of R-tree local indexes to ensure it fits one HDFS block
  - Fixed a bug with SpatialRecordReader#nextShapes that causes a few records to be skipped
  - Ensure that DistributedJoin produce answer pairs in the same ordering of input
  - Check input file existence in all operations