Alvaro Herrera <alvherre@commandprompt.com> wrote:
> > The basic concept is spooling new coming data, and merge the spool and
> > the existing indexes into a new index at the end of data loading. It is
> > 5-10 times faster than index insertion per-row, that is the way in 8.3.
>
> Please see
> http://thread.gmane.org/gmane.comp.db.postgresql.general/102370/focus=102901
Yeah, BEFORE INSERT FOR EACH ROW trigger is one of the problems.
I think it is enough to disallow bulkloading if there are any
BEFORE INSERT triggers. It is not a serious limitation because
DBA often disables triggers in bulkloading for performance.
>> You could work around this if the indexscan code knew to go search in the
>> list of pending insertions, but that's pretty ugly and possibly slow too.
I heard it is used in Falcon storage engine in MySQL, so it seems to be
not so unrealistic approach.
Regards,
---
ITAGAKI Takahiro
NTT Open Source Software Center
