Re: pgsql: If an index depends on no columns of its table, give it a

On Fri, 2007-11-09 at 11:28 -0500, Tom Lane wrote:
> Simon Riggs <simon@2ndquadrant.com> writes:
> > On Fri, 2007-11-09 at 04:05 -0500, Tom Lane wrote:
> >> That was my first reaction too, but the point about unique-index behavior
> >> refutes it.  Constraining a table to have at most one row is useful.
>
> > Sure is, and I've done it just a few days ago.
>
> > This SQL does it using standard syntax:
>
> >   create table foo (handle integer primary key check (handle = 1));
>
> That does not constrain the table to have only one row.  It constrains
> it to have only one value of the handle field (thereby making the field
> useless).

It works, sure you need another column to put data in.

> The fact that there are workarounds isn't a reason to not
> support the index option.

The above is not a workaround. It is the SQL Standard way of solving the
problem, so why support another non-standard way?

Constants in indexes are just a strangeness we don't need. Supporting
weird syntax because one person wants it has never been anything you've
advocated before, so I'm surprised to see that argument deployed here.

--
  Simon Riggs
  2ndQuadrant  http://www.2ndQuadrant.com