Re: Multi column range partition table

On 2017/07/03 14:00, Amit Langote wrote:
> On 2017/07/03 2:15, Dean Rasheed wrote:
>> On 30 June 2017 at 10:04, Ashutosh Bapat
>> <ashutosh(dot)bapat(at)enterprisedb(dot)com> wrote:
>>> On Fri, Jun 30, 2017 at 1:36 PM, Amit Langote
>>> <Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp> wrote:
>>>>
>>>> Alright, I spent some time implementing a patch to allow specifying
>>>> -infinity and +infinity in arbitrary ways. Of course, it prevents
>>>> nonsensical inputs with appropriate error messages.
>>>
>>> I don't think -infinity and +infinity are the right terms. For a
>>> string or character data type there is no -infinity and +infinity.
>>> Similarly for enums. We need to extend UNBOUNDED somehow to indicate
>>> the end of a given type in the given direction. I thought about
>>> UNBOUNDED LEFT/RIGHT but then whether LEFT indicates -ve side or +side
>>> would cause confusion. Also LEFT/RIGHT may work for a single
>>> dimensional datatype but not for multi-dimensional spaces. How about
>>> MINIMUM/MAXIMUM or UNBOUNDED MIN/MAX to indicate the extremities.
>>>
>>
>> Yes, I think you're right. Also, some datatypes include values that
>> are equal to +/-infinity, which would then behave differently from
>> unbounded as range bounds, so it wouldn't be a good idea to overload
>> that term.
>
> Agree with you both that using (+/-) infinity may not be a good idea after
> all.
>
>> My first thought was UNBOUNDED ABOVE/BELOW, because that matches the
>> terminology already in use of upper and lower bounds.
>
> I was starting to like the Ashutosh's suggested UNBOUNDED MIN/MAX syntax,
> but could you clarify your comment that ABOVE/BELOW is the terminology
> already in use of upper and lower bounds? I couldn't find ABOVE/BELOW in
> our existing syntax anywhere that uses the upper/lower bound notion, so
> was confused a little bit.
>
> Also, I assume UNBOUNDED ABOVE signifies positive infinity and vice versa.

Anyway, here's the revised version of the syntax patch that implements
ABOVE/BELOW extension to UNBOUNDED specification.

0001 is the patch that Dean posted [1] as a replacement for what I earlier
posted for simplifying range partition overlap check.

0002 is the UNBOUNDED syntax extension patch.

Thanks,
Amit

[1]
https://www.postgresql.org/message-id/CAEZATCVcBCBZsMcHj37TF%2BdcsjCtKZdZ_FAaJjaFMvfoXRqZMg%40mail.gmail.com