Create R-tree indexes

To use the R-tree secondary access method, you must first create an R-tree index on a column whose data type can be indexed by the R-tree access method.

You can create an R-tree index either before or after you insert data into the table. However, if you are loading large amounts of data into the table, you should create the R-tree index after you load the data. When you create an R-tree index on a loaded table, the generation of log records is suppressed, so you do not run out of log space. If, however, you create the index first and then load large amounts of data in a single transaction, you might run out of log space, which causes the transaction to abort.

In addition, if you use the bottom-up build method, described in later topics, to create the R-tree index after you have loaded the data, the size of the index is typically about two-thirds the size of the index built with a slower method. The R-tree access method uses bottom-up building when creating an R-tree index only when data currently exists in the table.