From: | Reg Me Please <regmeplease(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: COPY speedup |
Date: | 2007-12-13 21:31:42 |
Message-ID: | 200712132231.43056.regmeplease@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Il Thursday 13 December 2007 19:56:02 Tom Lane ha scritto:
> Reg Me Please <regmeplease(at)gmail(dot)com> writes:
> > In order to speed up the COPY ... FROM ... command, I've
> > disabled everything (primary key, not null, references, default and
> > indexes) in the table definition before doing the actual COPY.
> > Later I can restore them with ALTER TABLE ... and CREATE INDEX ...
> >
> > My question is: is all this necessary, or could I save some of them
> > (maybe just the DEFAULT) with no speed cost?
>
> Indexes and foreign key references are the only things that benefit
> from this treatment. DEFAULTs are irrelevant to a COPY, and simple
> constraints (NOT NULL and CHECK) are not any faster to verify later
> --- which makes dropping them slower, since you'll need an additional
> table scan to verify them when they're re-added.
>
> regards, tom lane
I'd suppose that foreign keys are to be "disabled" in order to speed things
up. Right?
--
Reg me, please!
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2007-12-13 21:41:55 | Re: [GENERAL] Slow PITR restore |
Previous Message | André Volpato | 2007-12-13 21:29:38 | Re: Hash join in 8.3 |