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: Column priority in table definition?

Re: Column priority in table definition?

From: Jerry Gitomer <jgitomer_at_erols.com>
Date: 2000/04/22
Message-ID: <20000422.5014600@noname.nodomain.nowhere>#1/1

        It's a nice theory, but when you consider that Oracle reads and writes blocks - not rows it starts falling apart. Even if it were true for the last row in the block if you update the first row in the block you still have to shift every other row in the block.

        You might as well make it easy on the humans and use logical groupings.

>>>>>>>>>>>>>>>>>> Original Message <<<<<<<<<<<<<<<<<<

On 4/21/00, 8:12:35 PM, Douglas Scott <dsscott_at_ev1.net> wrote regarding

Column priority in table definition?:

> I recently heard that when you create a table you should order the
> columns in the table in the following order to increase database
> efficiency. The order should be something like PK, mandatory fixed
> length, mandatory variable length, optional fixed length, optional
> variable length. PK columns should not be updated so they get to go
 to
> the head of the line. Fixed length, and mandatory columns get next
> priority because as they get updated they have the least effect on
> expanding and shrinking the row in which the data is written, so
 writing
> and future reads take less time. I really prefer to see columns
> logically grouped myself. Is the increase in efficiency enough to
> warrant this?
 

> Douglas Scott
Received on Sat Apr 22 2000 - 00:00:00 CDT

Original text of this message

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