Re: RelOptInfo -> Relation

On Tue, Feb 6, 2018 at 2:23 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> If we're going to have to change this at some point (and I bet we
>> are), I'd rather do it before people jam even more stuff into the
>> current system rather than wait until it gets totally out of hand.
>
> While I'm prepared to believe that this *could* be a problem, I repeat
> that you've offered no hard evidence that it actually *is* a problem,
> or might become one in the future.  We could easily expend a significant
> amount of effort here for no real return.

Sure.  The point of the email was to ask whether there was some
consideration of which I had not thought which makes this utterly
unthinkable; the answer at least so far seems to be "no".

I think it will be interesting to see what happens with the faster
partition pruning patch.  If it doesn't run into a problem with this,
then there's no urgent need to do anything here.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company