Hello there, I would like to get an idea, what (insert) performance I could expect when building a 3-dimensional (float) R*Tree.
I am using SQLite as a static lib, compiled with the following defines: "SQLITE_ENABLE_RTREE" "SQLITE_OMIT_DEPRECATED" "SQLITE_THREADSAFE=2" "SQLITE_MAX_EXPR_DEPTH=0" The test I performed was on Win7 64-bit. I am using an in-memory DB. The PC has lots of RAM. I have 3 tables: points consists of an implicit rowid and 3 real columns elemenst consists of 4 integer columns, referencing a points rowid each. Populating the above tables from an array of data goes with a speed of approx. 1.2 M elements per second. I am happy with this. I have a transaction wrapped around a for loop where I bind the parameters and execute the parametrized query, inserting one row at a time. Now, I created a table for spatial lookups: create virtual table boundingboxes using rtree (elemID, minX, maxX, minY, maxY, minZ, maxZ); I have a rather poor performance of 50 k inserts per second. The data to be inserted is precalculated and passed to the loop. I am using the same logic as above, so the loop is wrapped in a transaction and I use a parametrized statement. Is this the performance I can expect, compared to more than a million inserts into a "simple" table? Could I do something better? thanks in advance _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
