Re: Quote from comp.object
Date: 28 Feb 2007 12:01:13 -0800
On Feb 28, 10:25 am, "Kevin Kirkpatrick" <kvnkrkpt..._at_gmail.com>
> On Feb 28, 11:43 am, "Aloha Kakuikanu" <aloha.kakuik..._at_yahoo.com>
> > On Feb 28, 5:48 am, "-CELKO-" <jcelko..._at_earthlink.net> wrote:
> > > I think he missed the point of RDBMS versus hierarchical databases.
> > > Yes, a hierarchical database like IMS (which has more data in it than
> > > all the SQL enterprise products combined) devoted to one job will run
> > > really fast. You tune for that one job. But it will not handle any
> > > other jobs very well.
> > And not so much faster to be really concerned. Consider:
> > 1. Extracting a set of records from a single table by some criteria.
> > There has to be a B-tree walk in both cases. Performance is the same.
> > 2. Extracting all children of a parent record by foreign key index
> > versus extracting children by direct references. Logarithmic cost
> > versus constant which is also not a big deal.
> #2 assumes the database has not been optimized to address this need by
> physically storing the join rather than physically storing base tables
> separately (in which case you'd be looking at constant vs. constant).
Is physically storing join practical? It requires efficient materialized view update method, and there are concurrency concerns, right? Received on Wed Feb 28 2007 - 21:01:13 CET