Mohit Sindhwani
2014-06-19 04:27:39 UTC
Hello! We are using SQLite3 for storing geographical points that can be
queried using a bounding box (find everything that lies within this
box). Obviously, this query fits the capabilities of the RTree module
very well and it is a simple 2 dimensional search using an R-Tree that
has 5 columns.
However, since these are points that are stored in the table, x1=x2 and
y1=y2 when we do the insertion. As a former embedded systems engineer,
this feels like a waste since I can see that we are inserting exactly
the same value into the table.
INSERT into data_rtree(1000, 10, 5, 10, 5);
INSERT into data_rtree(1000, 17, 1, 17, 1);
and so on.
Is there a way that we could optimize the module so that we don't need
to store the same value twice? We are using this on a system with
constrained resources, so it helps to reduce the amount of storage space
we need for our database.
Thanks,
Mohit.
queried using a bounding box (find everything that lies within this
box). Obviously, this query fits the capabilities of the RTree module
very well and it is a simple 2 dimensional search using an R-Tree that
has 5 columns.
However, since these are points that are stored in the table, x1=x2 and
y1=y2 when we do the insertion. As a former embedded systems engineer,
this feels like a waste since I can see that we are inserting exactly
the same value into the table.
INSERT into data_rtree(1000, 10, 5, 10, 5);
INSERT into data_rtree(1000, 17, 1, 17, 1);
and so on.
Is there a way that we could optimize the module so that we don't need
to store the same value twice? We are using this on a system with
constrained resources, so it helps to reduce the amount of storage space
we need for our database.
Thanks,
Mohit.