btw, 9.4 should be wiser in case of rare+common terms,
thanks to GIN fast scan feature.
I'll look forward to it! We have a few other GIN indexes . . .
I don't want to misrepresent my impression of Postgres performance; the only
other case where I've made a significant improvement by tweaking w
As mentioned here and elsewhere (most recently in "How can I get the query
planner to use a bitmap index scap instead of an index scan ?" - 8 Mar
2014), estimation of the relative cost of text search operations using
GIN-indexed columns sometimes goes awry, particularly when there will be a
lar
