Re: Consequence of changes to CTE's in 12 - Mailing list pgsql-general

Michael Lewis <mlewis@entrata.com> writes:
> If you move the limit 1 to the first CTE, does it not give you the same
> behavior in both versions?

Not sure if that's exactly the same, but certainly adding a traditional
optimization fence (OFFSET 0) to the first CTE should do the trick.

            regards, tom lane