Date: Sun, 14 Jul 2024 00:05:45 +0200
From: "Peter J. Holzer" <hjp-pgsql@hjp.at>
To: pgsql-general@lists.postgresql.org
Subject: Re: Dropping column from big table
Message-ID: <20240713220545.cgjghaggksov3xkt@hjp.at>
Mail-Followup-To: pgsql-general@lists.postgresql.org
References: <CAD=mzVUGj2Zg0RY9YJwgDzAeghPjuT0FOnNs-aysTBGss87Yyw@mail.gmail.com>
 <f519a562-667c-4a96-9844-a2aaa42928b3@aklaver.com>
 <CAD=mzVUS8mFK42zmVUUC3JkiN=9QSUZ_Qxayd6=KJwbqdd0HzA@mail.gmail.com>
 <CANzqJaBAOzCjXaBFLa6cYweVNf5LbBvLYFkWx_hb3YezOP3quw@mail.gmail.com>
 <CAD=mzVXK+4tLMbENMecruPbk8cmeE4smD=ENUrwx5wh4N7hZhA@mail.gmail.com>
 <acc8b587497a65d38fafe4e0030e252eda78398f.camel@cybertec.at>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="t4huj3szjlklb37x"
Content-Disposition: inline
In-Reply-To: <acc8b587497a65d38fafe4e0030e252eda78398f.camel@cybertec.at>
Archived-At: <https://www.postgresql.org/message-id/20240713220545.cgjghaggksov3xkt%40hjp.at>
Precedence: bulk


--t4huj3szjlklb37x
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On 2024-07-11 10:06:47 +0200, Laurenz Albe wrote:
> On Thu, 2024-07-11 at 13:10 +0530, sud wrote:
> > Dropping will take it's own time for post vacuum however as you
> > rightly said, it won't be blocking which should be fine.=A0
>=20
> I am not certain if you understood this correctly.
>=20
> Dropping a column is fast, but doesn't reclaim the space.
> VACUUM won't block anything, but won't reclaim the space.
> VACUUM (FULL) will block everything, but will also not reclaim the space.
>=20
> You'd need to use a form of ALTER TABLE that rewrites the table,
> as indicated in the documentation.

Unfortunately the documentation indicates very little. It mentions that
the table will be rewritten with

* SET ACCESS METHOD
* a volatile DEFAULT
* changing the type of an existing column (unless binary coercible)

All three change something which you probably don't want to change.

The documentation also mentions some cases where the table is not
rewritten, so maybe some not explicitely mentioned options rewrite the
table, too.

I would especially expected ALTER TABLE ... CLUSTER to do this, but if
VACUUM FULL preserves the (former) content of dropped columns, maybe
CLUSTER does, too?

        hp

--=20
   _  | Peter J. Holzer    | Story must make more sense than reality.
|_|_) |                    |
| |   | hjp@hjp.at         |    -- Charles Stross, "Creative writing
__/   | http://www.hjp.at/ |       challenge!"

--t4huj3szjlklb37x
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=XBi1
-----END PGP SIGNATURE-----

--t4huj3szjlklb37x--