Just about as soon as I sent the above, I realized that it's unlikely to
make sense in the real world in a row-store. If the goal is to keep the
top-25 results and trim the rest, what happens when values are
added/modified/deleted? You now *have to go look at all of the data you
aren't caching in t
Looking for an index to support top-n searches, were n has a fixed maximum
Recently, I've been looking at strategies to handle top-n queries in
Postgres. In my current cases, we've got definition tables, and very large
related tables. Here's a stripped-down example, the real tables are much
wider.
