Re: truncating timestamps on arbitrary intervals

On 10.04.21 14:53, John Naylor wrote:
>
> On Sat, Apr 10, 2021 at 7:43 AM Peter Eisentraut
> <peter(dot)eisentraut(at)enterprisedb(dot)com
> <mailto:peter(dot)eisentraut(at)enterprisedb(dot)com>> wrote:
> >
> > On 30.03.21 18:06, John Naylor wrote:
> > > Currently, when the origin is after the input, the result is the
> > > timestamp at the end of the bin, rather than the beginning as expected.
> > > The attached puts the result consistently at the beginning of the bin.
> >
> > In the patch
> >
> > +   if (origin > timestamp && stride_usecs > 1)
> > +       tm_delta -= stride_usecs;
> >
> > is the condition stride_usecs > 1 really necessary?  My assessment is
> > that it's not, in which case it would be better to omit it.
>
> Without the condition, the case of 1 microsecond will fail to be a
> no-op. This case has no practical use, but it still must work correctly,
> just as date_trunc('microsecond', input) does.

Ah yes, the tests cover that. Committed.