Re: Is there any method to keep table in memory at startup
От | Vinay Jain |
---|---|
Тема | Re: Is there any method to keep table in memory at startup |
Дата | |
Msg-id | 409B864D.5040406@sarathi.ncst.ernet.in обсуждение исходный текст |
Ответ на | Re: Is there any method to keep table in memory at startup ("Andrew Dunstan" <andrew@dunslane.net>) |
Список | pgsql-hackers |
Andrew Dunstan wrote: >Vinay Jain said: > > >>Andrew Dunstan wrote: >> >> >> >>>If course it is not loaded each time. That would be insane. If not >>>preloaded it is loaded when first called in each process, and then >>>kept. It is never unloaded (except by the termination of the process >>>that loaded it). >>> >>>It seems you have been laboring under a misapprehension. >>> >>> >>> >>If this is the case than I can make connection to data base in starting >> of indchar.so file and close connection at end of it and it should >>work.... bingo!! >>it would solve my problem >> >> >> > >*sigh* > >You are not getting it. > >An immutable function MUST NOT DEPEND ON DATA IN THE DATABASE. Sorry to >shout but you really need to understand this. > >Use static C data, not database tables. In addition to it being pure, it >will also be enormously faster than getting data from the database. Yes it >means that if you want to change the lookup data you need to recompile >your C function library, and to redo any indexes etc. that depend on the >function. That's just the way it is, I'm afraid. > >cheers > >andrew >Hi > ya you are right but i wanted to escape from compilation and stuff.... I will think on declaring function stable rather than immutable and test performance.... for the time being i will follow your advice... thanks regards Vinay > > >---------------------------(end of broadcast)--------------------------- >TIP 8: explain analyze is your friend > > >
В списке pgsql-hackers по дате отправления: