I can’t imagine why you would think I would assume your question wasn’t serious, or why I wouldn’t try to supply a serious answer.

My answer, though, won’t sound entirely serious because it is the answer made famous by Douglas Adams: 42.
(See http://jonathanlewis.wordpress.com/2008/02/28/42/ for the serious justification).

Nigel has made some relevant comments, though, which I agree with. There is a point at which the data becomes so old that there is no point in holding it for comparison purposes. You might want to keep 13 months, for example, so that you could compare this Christmas with last Christmas – but you also have to worry about how you might handle creating reports across instance restarts.

In principle, so long as you remember to include the full, indexed, keys of the tables in your queries, you probably can’t do much harm (in terms of performance) when querying the tables – but I suppose you might still decide to limit the damage by reducing history to one or two snapshots per day and then compacting the tables at regular intervals.

Another alternative is to generate and archive one report per day once the data is over (say) 42 days and then delete the data.

This may give you some idea of the sort of thing you’re looking for:
http://jonathanlewis.wordpress.com/2007/03/07/analysing-statspack-4/#comment-31142

SQL> connect perfstat/perfstat
SQL> @$ORACLE_HOME/rdbms/admin/sppurge <– Purges a limited range of Snapshot IDs for a given database instance.

Yes. That is the sort of thing I was thinking of.
It is taking up a lot of space, but I am hesitant to get rid of it just because of the potential for analyzing and/or reporting on the high level stuff which could be used to show some sorts of totals or trends.

Thanks,

Robert.

This is a serious comment/question to which I very much hope you will reply with a serious answer.

I have been collecting statspack stats hourly at level 7 for about a year. I would like to keep about 3 years of these stats for analyzing trends, comparing historical data, etc. currently I have about 2 years of statspack data (the earlier data was level 5 and possibly not hourly).

Do you think this is excessive and/or a bad idea and/or a waste of time?

I would very much appreciate your comments on saving extensive statspack data for a given database.

Thank you,

Robert Wood.

tutu