From: | Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> |
---|---|
To: | Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, David Fetter <david(at)fetter(dot)org>, PG Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: GROUPING |
Date: | 2015-05-21 12:22:18 |
Message-ID: | CAEZATCUYmq3R74Ztkjvs1rC5LfN3juVyUNAM+t0f8KWeEuvMeA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 21 May 2015 at 09:20, Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> wrote:
>>>>>> "Dean" == Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> writes:
>
> >> Maybe INT8 would be a better choice than INT4? But I'm not sure
> >> there's any practical use-case for more than 30 grouping sets
> >> anyway. Keep in mind the actual output volume probably grows like
> >> 2^N.
>
> Dean> Actually using ROLLUP the output volume only grows linearly with
> Dean> N. I tend to think that having such a large number of grouping
> Dean> sets would be unlikely, however, it seems wrong to be putting an
> Dean> arbitrary limit on it that's significantly smaller than the
> Dean> number of columns allowed in a table.
>
> Limit on what exactly?
>
> Consider that in both MSSQL 2014 and Oracle 12 the limit on the number
> of arguments in a GROUPING() expression is ... 1.
>
Actually Oracle haven't quite followed the standard. They have 2
separate functions: GROUPING() which only allows 1 parameter, and
GROUPING_ID() which allows multiple parameters, and returns a bitmask
like our GROUPING() function. However, their GROUPING_ID() function
seems to return an arbitrary precision number and allows an arbitrary
number of parameters (well, I tested it up 70 to prove it wasn't a
64-bit number).
Regards,
Dean
From | Date | Subject | |
---|---|---|---|
Next Message | Heikki Linnakangas | 2015-05-21 12:35:13 | Archiving last incomplete segment as .partial issues |
Previous Message | Robert Haas | 2015-05-21 12:12:20 | Re: Change pg_cancel_*() to ignore current backend |