[GRASS-user] Large vector files

[Hamish]

Jonathan Greenberg wrote:
> I wonder (and I'm thinking out loud here) if there are ways to "tile"
> vector processes in an analogous (if not algorithmic) way to how we
> deal with massive raster datasets?

WRT point data: discrete points are stored as such. They know nothing
about the proximity of neighboring points without a lot of multi-
dimensional sorting & refiltering. There is no "gravity" to other
points. So the traditional method of iterating from the first point
until the last point is often the best. If points need to know about
their neighbors some sort of binning in to a grid (e.g. v.surf.rst
adaptive grid size) or sorting+threshold (v.surf.idw npoints=) has to be
done.

I can think of one case when it would be useful: point in polygon
extraction (doable with awk, but a pain).

A m.* module (frontend for awk) or C program using vector lib fns could
help.


Hamish