Re: [HACKERS] Range Partitioning behaviour - query - Mailing list pgsql-hackers

On 2017/02/24 10:38, David G. Johnston wrote:
> On Thu, Feb 23, 2017 at 6:17 PM, Amit Langote <Langote_Amit_f8@lab.ntt.co.jp
>> wrote:
> 
>> On 2017/02/24 8:38, Venkata B Nagothi wrote:
>>> On Thu, Feb 23, 2017 at 3:14 PM, Amit Langote wrote:
>>>> Upper bound of a range partition is an exclusive bound.  A note was
>> added
>>>> recently to the CREATE TABLE page to make this clear.
>>>>
>>>> https://www.postgresql.org/docs/devel/static/sql-createtable.html
>>>
>>>
>>> Thanks. Actually, my confusion was that the upper bound value would be
>>> included when "TO" clause is used in the syntax.
>>
>> Hmm, TO sounds like it implies inclusive.
>>
> 
> ​I think most common usage of the word ends up being inclusive but the word
> itself doesn't really care.​
> 
> Dictionary.com has a good example:
> 
> "We work from nine to five." - you leave at the beginning of the 5 o'clock
> hour (I'm going for casual usage here)

Thanks for that example.

One problem I've seen people mention is one of cognitive dissonance of
having to define partition_y2013 as FROM ('2013-01-01') TO ('2014-01-01'),
given that that's the only way to get what one needs. But we concluded
that that's a reasonable compromise.

> Since our implementation of ranges is half-open the usage here is
> consistent with that concept.  That it doesn't match BETWEEN is actually
> somewhat nice since you can use ranges for half-open and BETWEEN if you
> want to be concise with fully-closed endpoints.  But it is one more thing
> to remember.

Agreed.

Thanks,
Amit