Re: Hash partitioning.

On Tue, Jun 25, 2013 at 12:08:34PM -0400, Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> > I looked at predtest.c but I can't see how we accept >= and <= ranges,
> > but not CHECK (a % 16 == 3). It is the '%' operator? I am not sure why
> > the hashme() function is there. Wouldn't it work if hashme() was an
> > immutable function?
>
> No. Robert's description is exactly correct: it's a question of whether
> we can know that the semantics of function X have anything to do with
> the behavior of operator Y. In the case of something like CHECK (X >= 16)
> combined with WHERE X = 10, if the given = and >= operators belong to
> the same btree opclass family then we can assume that their semantics
> are compatible and then apply reasoning to show that these two clauses
> can't both be true for the same value of X. We can *not* use "X = 10"
> to reason about the behavior of anything that isn't in the = operator's
> btree opclass, because we don't assume that "=" means "absolutely
> identical for every purpose". And in fact it does not mean that for
> several pretty common datatypes (float being another example besides
> numeric).

OK, so it is really the index comparisons that we are using; makes
sense.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +