Re: [PGSQL 8.3.5] How to handle FKs with partitioning?

My idea is that DRI will help during the the JOINs I'll need to make later.

Creating a trigger to check the consistence would not help for that case,
unless my idea is wrong. In which case I'd follow the great Merlin's hint.

So the question is now: do DRI impact on JOINs efficiency? What'd be the gain?

The table in question should easily go 20+M rows, possibly up to 50+M a year.
The partitioning would ensure about 2M rows per partition and the trigger
should work accordingly to this (dynamic) schema.
So, along with the loss of efficiency due to the trigger I also would get some
other loss because of an external table needed for the partitioning.

On Friday December 19 2008 17:15:56 Merlin Moncure wrote:
> On Fri, Dec 19, 2008 at 6:04 AM, Reg Me Please <regmeplease(at)gmail(dot)com>
wrote:
> > Hi all.
> >
> > I need to implement something very similar to temporal table partitioning
> > as described in the documentation at chapter 5.9.
> >
> > My issues come from the fact that I have other tables that references
> > (FKs) to the table(s) to be partitioned. Those references are enforced by
> > means of DRI statements (REFERENCES ...).
> >
> > As the table containing the referenced data will not be a single table,
> > will I be forced to drop DRI?
> > The referencing table(s) don't need to be partitioned, though and have
> > also other FKs to other tables.
> >
> > Is there any other solution as I would keep DRI?
>
> Write a trigger.
>
> merlin

--
Fahrbahn ist ein graues Band
weisse Streifen, grüner Rand