Secondary indexes - response (11) by Shelley

Partitioned primary indexes are great for large tables where you consistently access the data by some type or range, ie day, week, month as an example. These will reduce full table scans especially if you only need one or two days out of several months or years of data,.
secondary indexes are used as another form of  access to the data. Ideally you would use a Pi for most access as it is the most efficient  but if you have additional columns that you use to access data from the table you might want to try a secondary index on column. Make sure you test each SI after you create them (and collect stats) to ensure they are being used. If they are not being used they are a waste of resources