Re: Performance on inserts

* Matthew Kirkwood <matthew(at)hairy(dot)beasts(dot)org> [000826 04:22] wrote:
> On Sat, 26 Aug 2000, Jules Bean wrote:
>
> > Is there any simple way for Pg to combine inserts into one bulk?
> > Specifically, their effect on the index files. It has always seemed
> > to me to be one of the (many) glaring flaws in SQL that the INSERT
> > statement only takes one row at a time.
>
> One of MySQL's little syntax abuses allows:
>
> INSERT INTO tab (col1, ..) VALUES (val1, ..), (val2, ..);
>
> which is nice for avoiding database round trips. It's one
> of the reasons that mysql can do a bulk import so quickly.

That would be an _extremely_ useful feature if it made a difference
in postgresql's insert speed.

>
> > But, using INSERT ... SELECT, I can imagine that it might be possible
> > to do 'bulk' index updating. so that scanning process is done once per
> > 'batch'.
>
> Logic for these two cases would be excellent.

We do this sometimes, works pretty nicely.

--
-Alfred Perlstein - [bright(at)wintelcom(dot)net|alfred(at)freebsd(dot)org]
"I have the heart of a child; I keep it in a jar on my desk."