Re: Have I found an interval arithmetic bug?

On Wed, Jul 28, 2021 at 08:42:31AM +0100, Dean Rasheed wrote:
> On Wed, 28 Jul 2021 at 00:08, John W Higgins <wishdev@gmail.com> wrote:
> >
> > It's nice to envision all forms of fancy calculations. But the fact is that
> >
> > '1.5 month'::interval * 2 != '3 month"::interval
> >
> 
> That's not exactly true. Even without the patch:
> 
> SELECT '1.5 month'::interval * 2 AS product,
>        '3 month'::interval AS expected,
>        justify_interval('1.5 month'::interval * 2) AS justified_product,
>        '1.5 month'::interval * 2 = '3 month'::interval AS equal;
> 
>     product     | expected | justified_product | equal
> ----------------+----------+-------------------+-------
>  2 mons 30 days | 3 mons   | 3 mons            | t
> (1 row)
> 
> So it's equal even without calling justify_interval() on the result.
> 
> FWIW, I remain of the opinion that the interval literal code should
> just spill down to lower units in all cases, just like the
> multiplication and division code, so that the results are consistent
> (barring floating point rounding errors) and explainable.

Here is a more minimal patch that doesn't change the spill-down units at
all, but merely documents it, and changes the spilldown to months to
round instead of truncate.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EDB                                      https://enterprisedb.com

  If only the physical world exists, free will is an illusion.