Strange query planner behavior

Поиск
Список
Период
Сортировка
От Qi Huang
Тема Strange query planner behavior
Дата
Msg-id BAY159-W50C8236FEDEEB74C315EC5A3020@phx.gbl
обсуждение исходный текст
Ответы Re: Strange query planner behavior  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-hackers
Hi, hackers 
   (I'm doing another project related to Postgres besides the gSOC, so this one is not for TABLESAMPLE)
   I have a query as follows :
select *
from affiliation a, author_aff_history his, author au
where a.aff_name = his.aff_name
and his.person_id = au.person_id;

I trace the RelOptInfo *rel in make_one_rel. In the pathlist, however, I find there are two duplicate paths(the latter two, both start with NESTPATH, and NEST on (b 2) and (b1) first, then with (b 3)). And there is another path that I think should be the optimal join tree(left-deep and no cartesian product), but I can't find it in the pathlist. The rel structure is attached in this email, the duplicate are in line 644 and 1321 respectively. 
I'm not sure whether I should ask this in the hacker list or other list, but just try here first.
Why is that so? Any suggestion?



Best Regards and Thanks
Huang Qi Victor
Computer Science of National University of Singapore
Вложения

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: psql bug
Следующее
От: Tom Lane
Дата:
Сообщение: Re: Strange query planner behavior