Re: Large table and buffer cache concerns

On 14 avr, 18:29, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk> wrote:
> <deva..._at_gmail.com> wrote in message
>
> As far as handling the data fast enough is concerned,
> your comments (and Joel's) about co-location are
> highly appropriate. You stated requirements sound like
> a perfect match for an Index OrganizedTablewith a
> primary key - picking up on your analogy - equivalent to
> (customer, video). All 1000 rows for a single customer
> would then fit into a single leaf block, and the random
> read problem ought to be minimised.

Woaw !

Jonathan, earlier I suggested to OP to investigate clustered tables...

While I have much experience with them, I have exactly zero experience with IOTs (were not available when physical design was performed)

Would you have an idea of the pro and con of one over the other, in the context of the OP requirements ?

Devalos, if you happen to benchmark both, would it be possible to post your results here ? That would be much appreciated.

Thanks to both

• Raoul