Re: planner complaints (was approve VKPts5 unsubscribe pgsql)

"Alexey V. Meledin" <avm@webclub.ru> writes:
> What mean "???" at (**).

It's probably a MATERIALIZE node.  The 6.5 explain-printer didn't have
code to recognize MATERIALIZE, so it prints like that.  (Fixed for 7.0.)

> Not optimal way I think

Possibly not.  It's hard to tell when you evidently have never vacuumed
your database.  Those tiny row counts and near-zero costs are presumably
bogus ... but if the planner thinks those tables are small, it's going
to generate plans accordingly.  You need to run a VACUUM, or preferably
VACUUM ANALYZE, to update the statistics the planner needs.

> 2. I've replace btree indexes on relation
> AND atd.ifs_data_id = def.ifs_data_id;
> with:
> create index itd_hash_data_id on ifs_tree_default USING HASH (ifs_data_id);
> create index iatd_hash_data_id on ifs_account_tree_data USING HASH (ifs_data_id);

Why would you do that?  The hash index method doesn't have any advantage
over btree that I can see, and it's got a lot of disadvantages.
        regards, tom lane