Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Understanding partitioning by hash

Re: Understanding partitioning by hash

From: Rick Denoire <100.17706_at_germanynet.de>
Date: Thu, 09 Oct 2003 00:59:18 +0200
Message-ID: <9859ov8js20iiufr1cqvbqp32qcqp76467@4ax.com>


"Howard J. Rogers" <hjr_at_dizwell.com> wrote:

>Everything will still work, but not exactly as before, because you're using
>a new feature. Yes, there are a number of new things to consider (such as
>an appropriate number of hash partitions, striking a balance between the
>benefits of partition elimination (the more partitios, the better) and poor
>data distribution (the 'right' number is important to determine, and more
>doesn't necessarily mean better).

So one hast to take a look at data distribution and adjust the number of partitions if necessary.

Eh, I like partitions! I even like partitioning by hash more than by range, specially when the range has no limit in one direction and repartitioning of the overflow partition would be necessary from time to time.

Thanks

Rick Denoire Received on Wed Oct 08 2003 - 17:59:18 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US