Re: "two time periods with only an endpoint in common do not overlap" ???

On 10/15/21 11:08 PM, Adrian Klaver wrote:
> On 10/15/21 19:42, Ron wrote:
>>
>> The numeric ranges 0-10 and 10-19 overlap, just as the time ranges
>> 00:01:00-00:00:02:00 overlaps 00:02:00-00:03:00.
>>
>> *It's the programmer's responsibility* to say what s/he really means, not
>> for "the system" to make that choice.
>>
>
> It is the programmers choice:
>
> --The canonical form.
> --No overlap
> select int4range(1,3), int4range(3, 5);
>  int4range | int4range
> -----------+-----------
>  [1,3)     | [3,5)
>
>  select int4range(1,3) && int4range(3, 5);
>  ?column?
> ----------
>  f
>
> --Making it overlap
> select int4range(1,3, '[]'), int4range(3, 5, '[]');
>  int4range | int4range
> -----------+-----------
>  [1,4)     | [3,6)
>
>  select int4range(1,3, '[]') && int4range(3, 5, '[]');
>  ?column?
> ----------
>  t
>
>
> There is no straight time range, you would have to use tsrange or
> tstzrange. The principle still holds though you can make ranges overlap or
> not depending on '[)' or '[]'.

OP refers to the OVERLAP operator (is it an operator), not the tsrange()
function.

--
Angular momentum makes the world go 'round.