Re: Early WIP/PoC for inlining CTEs

On Tue, Jul 24, 2018 at 6:10 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> My read of the concensus (in which I am in the majority, so I might be
> biased) is that we do want inlining to be the default. We were thinking
> that it'd be necessary to provide a way to force inlining on the SQL
> level for individual CTEs.

This is correct. Suggesting that we need syntax to disabling inlining
at the CTE level, and/or GUC to control the behavior (which I agree
should be defualted to inline). Something like
enable_cte_inline=true; I'm not very enthusiastic about explicitly
breaking intentionally introduced optimization fences and then forcing
people to inject our OFFSET 0 hack. This is just too unpleasant to
contemplate...what happens if we come up with a better implemntation
of OFFSET? yuck.

Thanks for providing this, CTE plan problems are a real bugaboo.

merlin