Enhancement: Permanent storage of R-trees?

Question

Thus, boost provides good spatial indexing capabilities in the form of R-trees. This is neat, but it is not yet possible to serialize the tree after it is assembled, am I mistaken?

The usual syntax out_archive <rtree does not work, because rtree does not have a serialize () member. In boost 1.57, there seems to be some experimental code for this, for example. /boost/geometry/index/detail/serialization.hpp, but it doesn’t actually compile!

So the first question is: does anyone know how to serialize an R-tree with boost?

If not, then my second question is: how are you going to constantly keep your index on disk to avoid having to rebuild it every time? (I have a dataset of 145M records, and it takes several hours to create the index, so I really don't want to build it more than once!)

Answer 1

• Packing Algorithm and Bulk Upload

  You can download packages (using the Packing algorithm).

  In addition, there are also algorithms that create an R-tree containing a number of objects. This method is called bulk loading and is performed using the packing algorithm [5] [6]. This method is faster and leads to R-trees with better internal structure. This means that query performance is increasing.

  _{[5] Leutenegger, Scott T .; Edgington, Jeffrey M .; Lopez, Mario A. (1997). STR: a simple and efficient algorithm for packing R-tree}

  _{[6] Garcia, Yvan J .; Lopez, Mario A .; Leutenegger, Scott T. (1997). Greedy loading algorithm for R-trees with bulk loading}

  More details: http://www.boost.org/doc/libs/1_57_0/libs/geometry/doc/html/geometry/spatial_indexes/introduction.html

• Using memory mapped files

  You can use a memory mapped file with custom allocators. This way you can use any view you want and it will be automatically saved

  More details: http://www.boost.org/doc/libs/1_57_0/libs/geometry/doc/html/geometry/spatial_indexes/rtree_examples/index_stored_in_mapped_file_using_boost_interprocess.html