Re: Performance on inserts - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Performance on inserts
Date
Msg-id 2963.967304748@sss.pgh.pa.us
Whole thread Raw
In response to Re: Performance on inserts  (Matthew Kirkwood <matthew@hairy.beasts.org>)
Responses Re: Performance on inserts
List pgsql-hackers
Matthew Kirkwood <matthew@hairy.beasts.org> writes:
> One of MySQL's little syntax abuses allows:
> INSERT INTO tab (col1, ..) VALUES (val1, ..), (val2, ..);

Actually, that's perfectly standard SQL92, just an item we haven't
got round to supporting yet.  (Until we do the fabled querytree
restructuring, it seems a lot harder than it is worth.)

COPY FROM stdin is definitely the fastest way of inserting data,
however, since you avoid a ton of parse/plan overhead that way.
Of course you also lose the ability to have column defaults
computed for you, etc ... there's no free lunch ...
        regards, tom lane


pgsql-hackers by date:

Previous
From: "Dominic J. Eidson"
Date:
Subject: Re: Access PostgreSQL server via SSL/Internet
Next
From: Tom Lane
Date:
Subject: Re: Proposal for supporting outer joins in 7.1