I would like to ask for advice!
Our directorate manages an Oracle Spatial database. We have about 2 millions of polygon geometries in a table.
In version 9.2 database we used Q-tree index, because the table was under heavy insert/update activity. But in newer versions of spatial databases the use of quadtree indexes is discouraged!
Does anyone have any experience in R-tree index in such kind of circumstances?
I have worked with Oracle databases that had 72 tables ranging from 2.5 to 10 million rows, all with SDO_GEOMETRY columns and all indexed with RTree indexes. 2 million is not going to be a problem.
One thing you may want to look into, if you expect your tables to grow, is partitioning your tables and partitioning your spatial indexes. I have very good experiences with partitioned spatial indexs ([url http://docs.oracle.com/cd/E11882_01/appdev.112/e11830/sdo_index_query.htm#i1005901]check the documentation for more info).