From: Tom Lane <tgl@sss.pgh.pa.us>
To: Ron <ronljohnsonjr@gmail.com>
cc: pgsql-admin <pgsql-admin@postgresql.org>
Subject: Re: pg_dump using anything other than custom and directory
In-reply-to: <93bf4376-6167-a89f-83c5-36aa0dadae2e@gmail.com>
References: <93bf4376-6167-a89f-83c5-36aa0dadae2e@gmail.com>
Comments: In-reply-to Ron <ronljohnsonjr@gmail.com>
	message dated "Fri, 12 Apr 2019 15:10:26 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <8579.1555110331.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 12 Apr 2019 19:05:31 -0400
Message-ID: <8580.1555110331@sss.pgh.pa.us>
Precedence: bulk

Ron <ronljohnsonjr@gmail.com> writes:
> In 2019 using supported versions of PostgreSQL, what practical use is th=
ere =

> to use the tar format, and -- other than migrating trivially sized datab=
ases =

> to other RDBMSs -- the plain format?

The historical argument for the tar format is that you can get your
data out of it with a standard Unix tool (tar, of course), rather than
having to depend on the availability of pg_restore.  Certainly there's
room to argue about how important that really is, but I don't think
the validity of the argument is much different than it was in 2001.

You need to be able to get a plain-text dump if you want to edit
the data or schema at all, which is a pretty common requirement.
However, as long as you're willing to assume the availability of
pg_restore, you can extract plain text from one of the other formats;
so this point isn't a reason not to make your dump in one of the
other formats to begin with.

			regards, tom lane