Point well taken, Niall. I had forgotten that we had used this aggregation method in a DW setting, and it is certainly apt in such a situation. Of course, I still recall a colleague scorning such usage. One problem you can have if they are not well-documented is that they can cause internal database performance issues if not monitored and tuned, and constantly verified as pertinent. What I'm getting at is that as business requirements change, or data shapes change, such reporting type MVs can become obsolete. I'd list this as another major "gotcha" in their use. _____ From: Niall Litchfield [mailto:niall.litchfield@xxxxxxxxx] Sent: Wednesday, December 24, 2008 4:46 AM To: mfontana@xxxxxxxxxxx Cc: Lee.Robertson@xxxxxxxxxx; oracle-l@xxxxxxxxxxxxx Subject: Re: Materialized Views On Tue, Dec 23, 2008 at 8:18 PM, Michael Fontana <mfontana@xxxxxxxxxxx> wrote: You need to know more about a feature before you can truly decide it's required. While it's allowed and there may be some valid reasons for it, often there are not. MVs are most often used to create distributed versions of data. T Hi Michael Complex aggregations, or rather complex aggregations that do not perform sufficiently raidly but which can be pre-calculated, are the other major reason for this feature in a reporting/dss type environment. To quote from the DataWarehousing guide One technique employed in data warehouses to improve performance is the creation of summaries. Summaries are special types of aggregate views that improve query execution times by precalculating expensive joins and aggregation operations prior to execution and storing the results in a table in the database. For example, you can create a summary table to contain the sums of sales by region and by product. So it rather sounded to me that Lee's developers reasons for using MVs looked to be not unreasonable. Having said all that, I;ve only ever really dealt with the remote view use for them, and you are absolutely right about the detection of out of date snapshots in that context. -- Niall Litchfield Oracle DBA http://www.orawin.info