2010/3/30 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
> I don't recall any operations at transaction start for such tables,
> but there may be some.
Both in СommitTransaction(void) and PrepareTransaction(void) we can
see PreCommit_on_commit_actions() call;
oids_to_truncate = lappend_oid(oids_to_truncate, oc->relid);
As a my opinion, the same action taked place at transaction start and commit.
To truncate rows of any temp tables, both changed and unchanged(!)
after transaction looks as little reinsurance. Well.
But why do the same action _before_ any transaction?
In response to
pgsql-performance by date
|Next:||From: Matthew Wakeling||Date: 2010-03-31 09:29:58|
|Subject: Re: experiments in query optimization|
|Previous:||From: David Wilson||Date: 2010-03-31 04:37:40|
|Subject: Re: query has huge variance in execution times|