On 2/26/15 3:22 AM, Andres Freund wrote:
> On 2015-02-26 02:20:21 -0600, Jim Nasby wrote:
>> The reason I'd like to do this with partitioning vs plain inheritance is
>> presumably as we build out partitioning we'll get very useful things like
>> the ability to have FKs to properly partitioned tables. Insert tuple routing
>> could also be useful.
>
> The problem there imo isn't so much inheritance, but lack of working
> unique checks across partitions. That's something we can implement
> independent of this, it's just not trivial.
There's been discussion of allowing for uniqueness when we can guarantee
no overlap between partitions, and the partition key is part of the
unique constraint. That's the particular use case I was thinking of.
I suspect there's other partitioning features that would be useful in a
generic inheritance setup as well; that's why I'd love to see both
features work together... but I fear there's enough work to get there
that it may not happen, and I don't want us to accidentally start mixing
the two and have users start relying on it.
--
Jim Nasby, Data Architect, Blue Treble Consulting
Data in Trouble? Get it in Treble! http://BlueTreble.com
