thanks , I can change the sql . Needed a little business modify in my application .but i don't know why in high load time I Executing the sql
,it cost 200-300ms . but the log record it cost 2000ms
log message:
duration: 2042.493 ms execute <unnamed>: select o_count from tshow.p_hpart_liveingroomlist_count($1)","parameters: $1 = '46'
657985552@xxxxxx
发件人： John R Pierce发送时间： 2015-10-28 16:28收件人： pgsql-general主题： Re: Re: Re: [GENERAL] 回复: postgres cpu 100% need helpOn 10/28/2015 12:12 AM, 657985552@xxxxxx wrote:
Thank you for your reply.
tshow=> explain (analyze, buffers) select count(t.*) from tshow.res_room_weight t,tshow.res_room_info r
tshow-> where t.subcatlg_id=46
tshow-> and t.roomid = r.actorid
tshow-> and r.levels>=0;
http://explain.depesz.com/s/oKN5
shows where the time is spent in that query, but as you say, that one the timing was fine. there's a lot of time being spent in a seq_scan of res_room_info.
I would have written that query as the inner join it is, but that probably won't effect how it's executed.
select ... from tshow.res_room_weight t join tshow.res_room_info r on (t.roomid = r.actorid) where t.subcatlg_id=46 and r.levels>=0;
-- john r pierce, recycling bits in santa cruz
