On 1/5/17, Vitaly Burovoy <vitaly(dot)burovoy(at)gmail(dot)com> wrote:
> On 1/4/17, Pantelis Theodosiou <ypercube(at)gmail(dot)com> wrote:
>> On Wed, Jan 4, 2017 at 3:03 PM, <web+postgresql(at)modin(dot)io> wrote:
>>
>>> The following bug has been logged on the website:
>>>
>>> Bug reference: 14486
>>> Logged by: Per Modin
>>> Email address: web+postgresql(at)modin(dot)io
>>> PostgreSQL version: 9.6.1
>>> Operating system: Linux 303a92173594 4.8.15-moby #1 SMP Sat Dec 17 0
>>> Description:
>>>
>>> Found this bug in 9.4.8, tried it in docker towards psql 9.6.1 and it's
>>> in
>>> there as well. A minimum working example would be as follows:
>>>
>>> ```
>>> postgres=# CREATE TABLE tbl AS SELECT 9223372036854 * interval '1
>>> second'
>>> col; TABLE tbl;
>>> SELECT 1
>>> ERROR: interval out of range
>>> ```
>>>
>>> ```
>>> postgres=# SELECT count(*) FROM tbl;
>>> count
>>> -------
>>> 1
>>> (1 row)
>>> ```
>>>
>>> It seems that inserting and retrieving data have different constraints.
>>> As
>>> you
>>> can see from query 2, the data still gets inserted.
>>>
>>> ```
>>> postgres=# select version();
>>> version
>>> ------------------------------------------------------------
>>> ------------------------------
>>> PostgreSQL 9.6.1 on x86_64-pc-linux-gnu, compiled by gcc (Debian
>>> 4.9.2-10)
>>> 4.9.2, 64-bit
>>> (1 row)
>>> ```
>>>
>>> Best regards,
>>> Per Modin
>>>
>>>
>> And these work without error:
>>
>> postgres=# select col - 9223372036854 * interval '1 second' from tbl ;
>> ?column?
>> ----------
>> 00:00:00
>> (1 row)
>>
>> postgres=# select col from xx where col < interval '100000 year' ;
>> col
>> -----
>> (0 rows)
>>
>
> Yes, it is a bug, but it is not a constraint, it is just different
> internal checks.
> Moreover even if a function does not raise an error, output could be wrong
> (pay attention to the duplicated '-' sign)
> postgres=# SELECT '-2147483647:59:59'::interval - '1s'::interval;
> ?column?
> --------------------
> --2147483648:00:00
> (1 row)
>
> I've written a patch which fixes that bug (in attachment).
> Should it be registered in the CF?
Oops. Forgot to attach the patch. Fixed.
--
Best regards,
Vitaly Burovoy