Hi, i am now using the rtree extension in my project, yet still see some slowdown for some scenarios (especially for searches in narrow/small rectangles).
I was therefore wondering if... - one can somehow index the fields of an rtree table, and if so how? - if there could be any other information in an rtree table other than the integer ID and then the real values for the rectangles. The reason i am asking for this is because this yields to aways the same setup - an rtree lookup table, which ndexes back to the real information via its ID. I would eventually like to ease my normal data table for the entries in the rtree table (in my case, i?d like to purge the longitude and latitude values that are used in the rtree table from the "normal" data table, as they are redundant in there), yet this yields to slower queries. Usually, my rtree is used in queries like select * from citydatabase where id in (select id from rtree where longitude_min>XX and longitude_max<YY and latitude_min>XY and latitude_max<XZ) Effectively, these are two queries, and the second one may results in a big subset (e.g. for big rectangle, like views from a high altitude on a 3d globe). I was also wondering if there is any way to get SQlite to cache the results of a previous query - since many of my requests are based on a Zoom in or zoom out in a given rectangle, it would probably be wise to cache an initial query with its results and use this for subsequent queries when zooming in for instance. Is there any function that would allow for tests like "new rect is inside older rect" ? Thanks for any insights, Christophe Leske _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
