I think Solr would handle this better personally, and you'd get
full-text search as an added bonus! :) But, of course, it is
advisable to give it a try and see. Solr is quite easy to get
rolling with, so it'd be well worth a try.
The one part I am most worried about is that solr would start doing
approximations when the amount of data grows big, like in the 500k
range, or even once I reach the 1M range.
Do you know if there is a way to make sure that all facet counts are
exact when dealing with this many records, and is it feasible to do
this, ie not making it too slow by forcing the exact counts?
solr does not do approximations. Faceting with large indexes (500K is
not particularly large) just requires RAM for reasonable performance.
Give it a try, and see what you think.
ryan
