On 4 December 2012 20:38, Jeff Davis <pgsql(at)j-davis(dot)com> wrote:
> Even if it is, I would
> think that a sufficiently narrow case -- such as CTAS outside of a
> transaction block -- would be safe
CTAS would be safe to do that with. CLUSTER and VACUUM FULL are already done.
Outside of a transaction block would be automatic.
Inside of a transaction block we could use a parameter called
initially_frozen. This would act same as FREEZE on COPY. Parameter
would not be recorded into the reloptions field on pg_class, since it
is a one-time option and has no meaning after end of xact.
> along with some slightly broader
> cases (like BEGIN; CREATE TABLE; INSERT/COPY).
Historically we've chosen not to allow bulk insert options on INSERT,
so that is a separate discussion.
Simon Riggs http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
In response to
pgsql-hackers by date
|Next:||From: Robert Haas||Date: 2012-12-07 16:41:03|
|Subject: Re: ALTER TABLE ... NOREWRITE option|
|Previous:||From: Bruce Momjian||Date: 2012-12-07 16:08:28|
|Subject: Re: pg_upgrade problem with invalid indexes|