Re: psql enhancement idea

From: "Uwe C(dot) Schroeder" <uwe(at)oss4u(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: psql enhancement idea
Date: 2004-10-21 22:00:20
Message-ID: 200410211500.20625.uwe@oss4u.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 21 October 2004 12:40 pm, Tom Lane wrote:
> "Uwe C. Schroeder" <uwe(at)oss4u(dot)com> writes:
> > I'm just doing a load of database transfers using pg_dump. Because I
> > changed the schema of quite some tables I have to dump full column
> > inserts.
>
> I don't think that conclusion follows from that premise. In recent
> pg_dump versions (any that use a column list with COPY, which I think is
> 7.3 or later) there is no fundamental disadvantage to using COPY; it
> should be semantically equivalent to INSERT-with-column-list commands.
>
> The only reason I can see for using the INSERT option anymore is if you
> think that some of the row insertions might fail, and you don't want
> that to stop the load altogether. If that's what you want, then you'd
> certainly not want to wrap the load in BEGIN/END, so there doesn't seem
> to be much point in adding an option to do it.
>
> regards, tom lane

The reason is that I made changes to the schema, i.e. changed the ordinal
position of columns and added some columns. Since the positioning of the
columns isn't the same anymore a copy will fail. Column Inserts will leave
the new columns untouched (set them to NULL or whatever the default) but can
handle the change in order of the columns (ok, the order thing is purely
cosmetic...).
I see your point that you don't want to fail the whole import when using full
inserts - which would happen in a transaction block. My point was basically
that I assume all inserts will succeed, however I can't use copy because of
the changes but doing it in a transaction will increase the insert speed by
at least 10 (chewing up memory certainly).

As a remark: normally I make pure data backups as copy and as full column
inserts. In case of a recovery I then can easily switch to single inserts if
the copy fails - eats a bit more diskspace but increases security (besides
the slony replication I'm running on top of that...)

Uwe

- --
Open Source Solutions 4U, LLC 2570 Fleetwood Drive
Phone: +1 650 872 2425 San Bruno, CA 94066
Cell: +1 650 302 2405 United States
Fax: +1 650 872 2417
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFBeDF0jqGXBvRToM4RAmrkAKCoRXnlmnRpp0DI5/cFeCyx97nm1QCfWJ/A
8zSHT4NXzG8LG7GFqpPGSbE=
=c99W
-----END PGP SIGNATURE-----

In response to

Responses

Browse pgsql-admin by date

  From Date Subject
Next Message Tom Lane 2004-10-21 22:03:16 Re: psql enhancement idea
Previous Message Gaetano Mendola 2004-10-21 21:49:35 Re: replication using WAL archives