to_char(interval) issues

Based on this discussion I have added these TODO items:

* Prevent to_char() on interval from returning meaningless values

For example, to_char('1 month', 'mon') is meaningless. Basically,
most date-related parameters to to_char() are meaningless for
intervals because interval is not anchored to a date.

* Allow to_char() on interval values to accumulate the highest unit
requested

o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
o to_char(INTERVAL '3 years 5 months','MM') => 41

Some special format flag would be required to request such
accumulation. Such functionality could also be added to EXTRACT.
Prevent accumulation that crosses the month/day boundary because of
the uneven number of days in a month.

---------------------------------------------------------------------------

Karel Zak wrote:
> On Sun, 2005-03-27 at 11:43 -0800, Josh Berkus wrote:
> > Tom, Karel,
> >
> > > Hmm, if we want to support conversion like:
> > > '43 hours 20 minutes' --> 'MI min'
> > > how we should work with calendar INTERVAL units? For example 'month'?
> > > '1 month 1 day' --> 'D days'
> > > I think answer should be error message: "missing calendar unit 'month'
> > > in output format"
> >
> > Actually, there's a pretty well-defined boundary within interval types:
> > year.month | day.hour.minute.second.millesecond
>
> Yes.
>
> > This subtype boundary of intervals is even defined in the SQL spec.
> >
> > > Surely not. to_char for timestamps doesn't require that you output
> > > every field of the value, and it shouldn't require that for intervals
> > > either.
> >
> > That's an invalid comparison. There is no logical way to "roll up" timestamps
> > into larger/smaller subtypes. There is with intervals.
>
> Agree. There is two possible way how you can convert it:
>
> a) extract and convert
>
> '1h 10min 30s' --- 'MI "min"' ---> '10 min'
>
> b) hold the interval and convert it to defined units
>
> '1h 10min 30s' --- 'MI "min"' ---> '70.5 min'
>
> > If you're arguing that this kink in the *useful* behavior of interval-->text
> > conversion is confusingly inconsistent with what to_char does with other data
> > types, and we should call the function something else, then I could
> > potentially buy that (assuming that others agree). However, our proprietary
> > functions are about being *useful*, not adhering to some unwritten de-facto
> > standard. And I am, as someone who uses intervals heavily in applications,
> > trying to define what the useful behaviour will be from a user's perspective.
>
> I agree with Josh that for interval is more useful second way where
> result from conversion is still useful interval.
>
> There is no problem implement both, to_char() stuff already supports
> global options and I can add for INTERVAL option 'EX' as extract.
>
> a) to_char('1h 10min 30s', 'EXMI "min"') -> '10 min'
> b) to_char('1h 10min 30s', 'MI "min"') -> '70.5 min'
>
>
> BTW, for numbers to_char() disable extraction:
>
> test=# select to_char(123.4::float, '.999');
> to_char
> ---------
> .###
>
> the result is not '.4'. I think important is always tradition how people
> work with selected datetype. For TIMESTAMP is it common that you work
> with extraction from full date/time description, but it's unusual for
> numbers and I think for INTERVALs too.
>
> Karel
>
> --
> Karel Zak <zakkr(at)zf(dot)jcu(dot)cz>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 7: don't forget to increase your free space map settings
>

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073