On 3/30/23 20:01, Peter Smith wrote:
> For example, Just imagine if logic could be made smarter to recognize
> that since there was already the 'part_def' being subscribed so it
> should NOT use the default 'copy_data=true' when the REFRESH launches
> the ancestor table 'part'...
>
> Even if that logic was implemented, I have a feeling you could *still*
> run into problems if the 'part' table was made of multiple partitions.
> I think you might get to a situation where you DO want some partition
> data copied (because you did not have it yet but now you are
> subscribing to the root you want it) while at the same time, you DON'T
> want to get duplicated data from other partitions (because you already
> knew about those ones -- like your example does).
Hm, okay. My interest here is mainly because my logical-roots proposal
generalizes the problem (and therefore makes it worse).
For what it's worth, that patchset introduces the ability for the
subscriber to sync multiple tables into one. I wonder if that could be
used somehow to help fix this problem too?
> At least, we need to check there are sufficient "BE CAREFUL" warnings
> in the documentation for scenarios like this.
Agreed. These are sharp edges.
Thanks,
--Jacob