Re: PCTUSED and PCTFREE

        Answers inline

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

On 4/17/00, 2:27:02 PM, mkb <mkb125NOmkSPAM_at_yahoo.com.invalid> wrote regarding PCTUSED and PCTFREE:

> Folks,
 

> Here's a scneario I have:
> I have a table that never get updated or deleted but regularly
> is truncated and inserted.
 

> If I set pctfree 0 and pctused 100 will there be chaining?

        Not if the row size is less than the block size.

> I believe that their will be since I am forcing the block to be
> completey used (pctused 100). Is this assumption correct?

        Not quite. Oracle will store as many full rows as it can in the block. So unless your data fills all of the available space you can have some unused space.

(my
> thinking here is that say my block size 8k and avg row size is
> 400 bytes. After overhead, I have a little less than 8k.
> Eventually, I'll end up in a situation where I have less than
> 400 bytes free in the block, say 200 bytes. I insert another
> row. Will Oracle place half the row in this block and chain the
> other half?).

        No.

> Would it better to go pctfree 0 pctused 95 or pctfree 5 pctused
> 90/95?

        If you really aren't going to do any updating at all I don't think it will help. I have read that PCTFREE + PCTUSED should be less than 100, but I have never seen any explanation as to why they shouldn't be exactly 100.

> Any ideas?
 

> thx
 

> mkb

> * Sent from RemarQ http://www.remarq.com The Internet's Discussion
 Network *
> The fastest and easiest way to search and participate in Usenet -
Free! Received on Tue Apr 18 2000 - 00:00:00 CDT