Re: [HACKERS] Postgres Speed or lack thereof

On Sun, 17 Jan 1999, Tom Lane wrote:

> Tom <tom(at)sdf(dot)com> writes:
> > On Sun, 17 Jan 1999, Tom Lane wrote:
> >> I tried this myself and found that wrapping BEGIN/END around a series of
> >> INSERT statements didn't make much difference at all.
>
> > Using what API?
>
> Sorry, I neglected to specify that it was psql (being driven by
> hand-trimmed pg_dump scripts).

This also wouldn't you give you any benefit from "prepare" that would
speed up repeated executions of the same statement. PostgreSQL does allow
statements to be prepared right? If it doesn't, chances are COPY has a
way of doing it. It saves a lot of parsing and planning time.

> > Some APIs control autocommit for you, so execing
> > "BEGIN" and "END" commands may not do anything.
>
> That's not the case for psql. If it were, I would have measured no
> difference in speed, rather than a small difference.
>
> regards, tom lane

Tom