The key slowdown is in bugmessage<->message establishment. The following should let us evaluate dropping a table in determining that on a staging server:
begin;
alter table bugmessage add column owner int;
update bugmessage set owner=message.owner from message where message.id=bugmessage.message;
create index bugmessage__owner__idx on bugmessage using btree(owner);
create index bugmessage__owner__index__idx on bugmessage using btree(owner, index);
create index bugmessage__commentor__idx on bugmessage using btree(owner) where index > 0;
commit;
The key slowdown is in bugmessage< ->message establishment. The following should let us evaluate dropping a table in determining that on a staging server:
begin; id=bugmessage. message; _owner_ _idx on bugmessage using btree(owner); _owner_ _index_ _idx on bugmessage using btree(owner, index); _commentor_ _idx on bugmessage using btree(owner) where index > 0;
alter table bugmessage add column owner int;
update bugmessage set owner=message.owner from message where message.
create index bugmessage_
create index bugmessage_
create index bugmessage_
commit;