On Wed, Feb 08, 2012 at 09:37:01AM -0500, Robert Haas wrote:
> On Tue, Feb 7, 2012 at 12:43 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> > I've committed the numeric and varbit patches and will look at the
> > temporal one next.
>
> Committed, after changing the OIDs so they don't conflict.
>
> Here's one more case for you to ponder:
>
> rhaas=# create table noah (i interval day);
> CREATE TABLE
> rhaas=# alter table noah alter column i set data type interval second(3);
> DEBUG: rewriting table "noah"
> ALTER TABLE
>
> Do we really need a rewrite in that case? The code acts like the
> interval range and precision are separate beasts, but is that really
> true?
The code has a thinko; a given interval typmod ultimately implies a single
point from which we truncate rightward. The precision only matters if the
range covers SECOND. Thanks; the attached patch improves this.