On Thu, Jul 30, 2020 at 6:39 PM James Coleman <jtc...@gmail.com> wrote: > I very much do not like this approach, and I think it's actually > fundamentally wrong, at least for the memory check. Quicksort is not the only > option that uses memory. For now, there's only one option that spills to disk > (external merge sort), but there's no reason it has to remain that way.
I wouldn't be surprised if it was possible to get SORT_TYPE_EXTERNAL_SORT even today (though I'm not sure if that's truly possible). That will happen for a regular sort node if we require randomAccess to the sort, and it happens to spill -- we can randomly access the final tape, but cannot do a final on-the-fly merge. Say for a merge join. -- Peter Geoghegan
