Re: partitioned tables referenced by FKs

Hi,

On 3/29/19 11:22 AM, Alvaro Herrera wrote:
> On 2019-Mar-29, Jesper Pedersen wrote:
>
>> Could expand a bit on the change to DEPENDENCY_INTERNAL instead of
>> DEPENDENCY_PARTITION_PRI / DEPENDENCY_PARTITION_SEC ?
>
> The PARTITION dependencies work in a way that doesn't do what we want.
> Admittedly, neither does INTERNAL, but at least it's less bad.
>
>> If you run "DROP TABLE t2_p32 CASCADE" the foreign key constraint is removed
>> from all of t1.
>
> Yes. CASCADE is always a dangerous tool; if you run the DROP partition
> without cascade, it explicitly lists that the constraint is going to be
> dropped.
>
> If you get in the habit of added CASCADE to all your drops, you're going
> to lose data pretty quickly. In this case, no data is lost, only a
> constraint.
>

Thanks !

Maybe the "(" / ")" in the CASCADE description should be removed from
ref/drop_table.sgml as part of this patch.

Should catalogs.sgml be updated for this case ?

Best regards,
Jesper