Re: single bit integer (TINYINT) revisited for 8.5

On Fri, 2009-07-03 at 13:38 -0400, Robert Treat wrote:
> On Thursday 02 July 2009 12:40:49 Simon Riggs wrote:
> > On Wed, 2009-07-01 at 11:19 -0400, Caleb Cushing wrote:
> > > A couple of times I've been told "you don't need tinyint, use boolean"
> > > which is not true, several projects I've worked on I've needed and
> > > integer field that supports number within a small range 0-5 1-10 1-100
> > > or something similar. I end up using smallint but it's range is huge
> > > for the actual requirements.
> >
> > Completely agree.
> >
>
> Blech. More often than not, I find people using all these granular types to be
> nothing more than premature optimization. And if you really do need a single
> byte type, you can use "char" (though again I'm not a big fan of that)

I agree that many optimizations are used inappropriately. Another reason
for making it an add-on module.

I'm aware of "char" and it doesn't do all I would wish.

> > I'm most or the way through working on this as an add-on module, rather
> > than a new datatype in core. I don't see much reason to include it in
> > core: its not an SQL standard datatype, it complicates catalog entries
> > and most people don't need or want it.
> >
>
> That's too bad. I'd much rather see someone implement something closer to
> Oracle's number type.

Please explain what you mean?

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support