On Tue, 15 Jul 2025 at 04:49, Robert Treat <rob@xzilla.net> wrote:
>
> On Mon, Jul 14, 2025 at 3:21 AM Dean Rasheed <dean.a.rasheed@gmail.com> wrote:
> >
> > But it's completely trivial to emulate random(min_date, max_date), just by doing
> >
> > min_date + random(0, max_date - min_date)
> >
> > Is it really worth adding a core function for that?
>
> I feel like this is a very similar argument against what was
> ultimately the addition of timestamp based generate_series functions,
> and similarly I think adding these in would be a rather useful
> improvement for users, though like generate_series, we don't need to
> hit every different data type (no one should ever generate a random
> timetz for instance).
Right, and for generate_series() we didn't add a variant for type
date. In fact, we explicitly documented how a sequence of dates can be
generated using the date-plus-integer operator.
In fact, I think generate_series() sets a good precedent, and I could
get behind a proposal that adds new random() functions with the same
signatures as the timestamp-based generate_series() functions. In
particular, I think that it's quite likely that if I wanted a random
timestamp in some range, I would want some specified precision, like
say 'hour' or 'day', and not a timestamp with some random number of
microseconds.
(In a similar vein, it might be useful to add an optional "step"
parameter to the random integer/numeric functions, but that's really a
separate proposal.)
Regards,
Dean
