On 9/13/18, Jens Alfke <j...@mooseyard.com> wrote: > The geopoly docs says that only the geopoly_overlap(P1,P2) and > geopoly_within(P1,P2) functions are optimized to use the R-tree index when > used as a WHERE test; so that means the geopoly_contains_point function is > not. > > This implies that, if I want to query for polygons containing a given point, > I should avoid using the obvious function geopoly_contains_point; instead I > should outset the point (by some epsilon) into a small rectangle, convert > that to a polygon, and use one of the two optimized functions. Correct?
Correct. > > Which prompts the question of whether I can use epsilon=0; i.e. how do the > overlap and within functions behave if one polygon is an empty (zero area) > rectangle? Will they work correctly, or always return false? I think a zero-area polygon will work. But, honestly, I don't remember writing a test case for that. (Probably, I should go back and add a few.) Please try it and see what happens :-) Do you have specific plans to use geopoly? Is this new extension meeting a specific need that you have? Or are you just curious? -- D. Richard Hipp d...@sqlite.org _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users