On Tue, 23 Jun 2009, Markus GRASS wrote:
My implementation is completely file based, also when creating or updating a vector. This comes obviously with a speed penalty because reading in memory is faster than reading from file. With all sorts of tricks and relying on the system to cache files, the file based spatial
Hi Markus, The last point above sounds interesting. What sort of tricks? Are you using memory-mapped files? My thinking would be to always read it from the file, and rely on the operating system as best as possible to not slow down performance too much (compared to holding the index in memory) for small sizes of indexes. I think that would be a seamless foolproof (from the user's point of view) approach that would scale very well with both increasing size of datasets and performance of computer hardware.
Paul _______________________________________________ grass-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-dev
