Re: logical changeset generation v3 - comparison to Postgres-R change set format

On 2013-01-13 12:44:44 -0500, Tom Lane wrote:
> Hannu Krosing <hannu@2ndQuadrant.com> writes:
> > How do people feel about adding a real sameness operator ?
> 
> Just begs the question of "what's sameness?"
> 
> In many places we consider a datatype's default btree equality operator
> to define sameness, but not all types provide a btree opclass (in
> particular, anything that hasn't got a sensible one-dimensional sort
> order will not).  And some do but it doesn't represent anything that
> anyone would want to consider "sameness" --- IIRC, some of the geometric
> types provide btree opclasses that sort by area.  Even for apparently
> simple types like float8 there are interesting questions like whether
> minus zero is the same as plus zero.
> 
> The messiness here is not just due to lack of a notation.

FWIW *I* (but others might) don't plan to support that case for now, it
just seems to be too messy for far too little benefit.

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services