On 2018/06/14 10:57, Michael Paquier wrote:
> On Wed, Jun 13, 2018 at 03:39:04PM +0200, Jürgen Strobel wrote:
>> I agree, and I imagine especially being able to use per-partition
>> sequences would be a common use case. That was my motivation when I
>> discovered it, and it was very counter intuitive to me.
>>
>> Thanks for the NULL violation patch btw.
>
> Please note that I have added this thread to the Open item page, under
> the older bug queue, and that a CF entry has been added so as we don't
> lost track of it:
>
> https://commitfest.postgresql.org/18/1671/
Thank you, Michael.
Just to be sure, you meant to add the open item for the NOT NULL violation
bug fix patch mainly or all of what appears to be wrong here? About
partition-specific defaults being ignored when inserting through the
parent, do you think we should consider it a bug? I could imagine fixing
the documentation to say that partition-specific defaults are only honored
if directly inserted into and ignored otherwise.
Thanks,
Amit