Re: [SQL] rule plan too big (fwd)
От | wieck@debis.com (Jan Wieck) |
---|---|
Тема | Re: [SQL] rule plan too big (fwd) |
Дата | |
Msg-id | m10paIL-0003kGC@orion.SAPserv.Hamburg.dsh.de обсуждение исходный текст |
Ответ на | rule plan too big (fwd) (Phil DiCorpo <pd3@wuerl.wustl.edu>) |
Ответы |
Re: [SQL] rule plan too big (fwd)
|
Список | pgsql-sql |
> I'm trying to create a relatively large view and I'm getting the > "DefineQueryRewrite: rule plan too big" error. Is this a bug in Postgres > that currently has no workaround, or is there something I can do. A > previous posting mentioned changing the blocksize. I thought I made this > change correctly, but it did not get rid of the error. Any help would be > much appreciated. Currently there is no other workaround than increasing the overall database blocksize until it fits. At the time I mentioned changing the rule plan (and qualification) storage to be able to split over tuples there was discussion about a general tuple split capability, which would have solved this problem silently. Unfortunately this general solution did not incarnate for v6.5. I've also played around with another data type which uses an internal format that is a simple, byte oriented lz compression. Using that on rule strings approximately compresses with rates of 40-70% (better on large rules == complex views). For now I'm planning to wait until v6.5 is released and then make a feature patch where the rule plans will be stored compressed. That will lower the problem (not entirely fixing it) and offer the new compressed type for other purposes too. ******************************************************* * WE STILL NEED THE GENERAL TUPLE SPLIT CAPABILITY!!! * ******************************************************* I vote for sticking this onto the TODO and make it a permanent show-stopper for version 7.0. NO v7.0 without unlimited tuple size! I know that this is about a year or so ahead :-) Jan -- #======================================================================# # It's easier to get forgiveness for being wrong than for being right. # # Let's break this rule - forgive me. # #========================================= wieck@debis.com (Jan Wieck) #
В списке pgsql-sql по дате отправления: