On Tue, 2006-01-17 at 21:43 +0000, Simon Riggs wrote:
> OK.... My interest was in expanding the role of HashAgg, which as Rod
> says can be used to avoid the sort, so the overlap between those ideas
> was low anyway.
Am I right in thinking that HashAgg would almost always be quicker than
SortAgg, even for large (> memory) aggregation sets? (Except where the
prior ordering has already been forced via an ORDER BY).
If that is so, then I will probably look to work on this sooner,
especially since we seem to have a clear design.
I'd originally viewed the spill-to-disk logic as a safety measure rather
than as a performance feature.
Best Regards, Simon Riggs