Stephen Frost <sfrost@snowman.net> writes:
> Entirely agree with this. Index skip scan is actually *ridiculously*
> useful in terms of an improvement, and we need to get the right people
> together to work on it and get it implemented. I'd love to see this
> done for v15, in particular. Who do we need to coordinate getting
> together to make it happen?
It sounds like Peter is willing to take point on the executor end
of things (b-tree in particular). If he can explain what a reasonable
cost model would look like, I'm willing to see about making that happen
in the planner.
regards, tom lane