An example using the libspatialindex library to bulkload points into an in-memory R*-Tree. This example loads points into an R-Tree using the bulkloading STR method. The example was adapted from here. It contains some unnecessary stuff such as using a custom-made class "Point.h" and assuming that each point belongs to a cluster.
- It implements a DataStream to handle our input.
- It does some text processing to convert the text data to a stream of SpatialIndex::RTree::Data objects.
- Each SpatialIndex::RTree::Data object is a region with two coordinates, low and high
- Since our data are points, each object in the stream is a region with low==high. See here for a discussion that clarifies that.
- It stores a cluster id (as a sequence of bytes) along with each region (i.e. point)
- Reads the points from a stream and inserts them directly into an R*-Tree using the STR bulkloading method
- It selects 600 random points and uses them as query points
- It queries the R-Tree and estimates the average time for retrieving a point. It also estimates the error of the measurements
- Since the application reads from a stream, it does not create unnecessary objects. However, the application momentarily uses twice as much memory as the size of the data. This is because the stream is a stream of regions and each region holds two coordinates (low and high that are basically the same coordinates, i.e. the point's coordinates). As soon as the building is done however, memory usage is equal to the size of the input data (see massif/12508.png).
- A sample file is provided
Clear instructions for installing the library can be found here
Assumes libspatialindex is properly installed
mkdir build
cd build
cmake ..
make
./SpatialIndexRtree sample.txt