Re: Allow parallel DISTINCT

On Wed, 18 Aug 2021 at 08:50, Zhihong Yu <zyu(at)yugabyte(dot)com> wrote:
> The patch is good from my point of view.

Thanks for the review. I looked over the patch again and the only
thing I adjusted was the order of the RESETs in the regression tests.

I left the " if (distinct_rel->pathlist == NIL)" ERROR case check so
that the ERROR is raised before we call the FDW function and hook
function to add more paths. Part of me thinks it should probably go
afterwards, but I didn't want to change the behaviour there. The
other part of me thinks that if you can't do distinct by sorting or
hashing then there's not much hope for the hook to add any paths
either.

I've pushed this to master now.

David