On 17 June 2010 20:51, J Chris Anderson <[email protected]> wrote: > > On Jun 17, 2010, at 10:49 AM, afters wrote: > > Is it correct that reductions spread up the b-tree only as high as needed > to > > satisfy the group-level demands? > > No. Because the group_level isn't known until query time, the index is the > same regardless of how you query it. So all reduce functions should be > designed to work properly without grouping. > > In case the updating of the view were made solely on read, would there still be a need to propagate reduce to the nodes higher than what was asked for?
Anyway, that's beside my scaling issue (and my area of expertise). I think a separate query for the text is appropriate and won't be too hard to implement. > Chris
