Re: Challenge: Partitioning is a wrong idea

HansF wrote:
> On Thu, 07 Apr 2005 11:05:49 -0700, mikharakiri_nospaum interested us
by
> writing:
>
> >
> > It is undeniable that partitioning concept introduces extra
> > complications. You have to be aware of many extra technicalities:
what
> > is partition prunning, what is partition wise join, etc.
>
> Funny thing about the above statement: the big benefit of
partitioning is
> from a developer's and user's perspective, they do not need to know
about
> partition pruning or partition wise join to get the related benefits.

Well, views are supposed to be such an abstraction. A DBA/Data architect is worried about physical layout for the relations, while the user just see relations as logical entities. It came as a great surprise to me that views somehow don't work in such scenario, and we have to introduce a new concept.

> So it's basically a case of moving the responsibility from the [get
it
> done fast at any cost] developers to the [the buck stops here] DBAs.
<G>

Hmm. I always wondered why I was never excited to be a DBA?

> Of course, things like: transportable tablespaces allowing the
buiding of
> just the lastest set of data and snapping that into the warehouse,
> requiring about 37 seconds of downtime in the warehouse;

Well, inserting into the normal table is slow, but you have no downtime at all!

> archiving data
> that should be moved from online to near-line in minutes; wiping only
> portions of data to be able to reload it;

As I mentioned already, there has yet to be a convincing argument why deletion performance is critical.

> data-oriented indexing rather
> than table-oriented indexing - these things are irrelevant.

I lost you here. What is "data-oriented indexing" vs "table-oriented indexing"? Received on Thu Apr 07 2005 - 14:42:19 CDT