check out SASH v2 on http://ashmasters.com/ash-simulation

 The old version v1 required the partioning option which required Enterprise Edition, so I re-wrote the collection code to

     1) install on windows as well as UNIX
     2) simplify the code
     3) use “poor mans partitioning”
Poor mans partitioning takes 7 ASH like tables and does a using “union all” to create the view v$active_session_history. This allows me to use truncates instead of deletes when purging data. Truncates are much cheaper than deletes and since alot of data is being collected the option of truncating is important. The collection package on the target inserts into a view on the repository called SASH. The view SASH points to one of 7 base tables SASH1 - SASH7. If there is no purge job running the the table base table will just keep filling up. If the optional job sash_repo.purge is started, then it will change which base table SASH points to (sash1 through sash7). Then every minute, hour or day (configurable) the job switches the view sash to point to the next base table, truncating the base table before the switch so that there is 7 minutes, 7 hours or 7 days of rolling data.