Reverse Indexes

All the examples on the ‘interweb’ that talk about situations where reverse indexes are useful, centre around inserting a large numbers of rows, in a short period of time, into a table where the primary key is a sequence number. This results in ‘hot spots’ where all the inserts need the same part of the index. Using a reverse index would spread the inserts out over the entire structure more evenly.

The downside is that a reverse index cannot be used for an index range scan. It does, however, work well with exact matches e.g. b_id = 1234567 and will full scan the index, if the CBO thinks it is appropriate.