> > I think you need to know which tables are involved and if they were > modified.
Of course. While creating a cache entry for a SELECT, we need to analyze it and extract tables involved in the SELECT. The information should be stored along with the cache entry. If any of the tables were modified, cache entries using the table must be removed. (these are already implemented in Pgpool-II's in memory query cache)
How do you handle tables hiding behind views? Also how does cached entries in pgpools know if some tables are modified without going thru pgpool, eg pgplsql or trigger or via psql directly?