From: Tom Lane <tgl@sss.pgh.pa.us>
To: Ron Johnson <ronljohnsonjr@gmail.com>
cc: "pgsql-general@postgresql.org" <pgsql-general@postgresql.org>
Subject: Re: Suggestion for Easier Cross-Database Query Handling in PostgreSQL
In-reply-to: 
 <CANzqJaBVAykTGq8j-gijJS-L0bK7duyO-k2xKgr8ZmO9mqxXMw@mail.gmail.com>
References: 
 <CAHWFTRjCy3U30whQJBw8z5_8boq=V9MzbtHMh6HZVP3o-OaGxg@mail.gmail.com>
 <F93CFB3A-0269-4261-804C-0CBAB5DF089D@elevarq.com>
 <CANzqJaBVAykTGq8j-gijJS-L0bK7duyO-k2xKgr8ZmO9mqxXMw@mail.gmail.com>
Comments: In-reply-to Ron Johnson <ronljohnsonjr@gmail.com>
	message dated "Fri, 15 May 2026 18:56:46 -0400"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1439461.1778887378.1@sss.pgh.pa.us>
Date: Fri, 15 May 2026 19:22:58 -0400
Message-ID: <1439462.1778887378@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1439462.1778887378%40sss.pgh.pa.us>
Precedence: bulk

Ron Johnson <ronljohnsonjr@gmail.com> writes:
> Why doesn't the shared OID space make cross-database queries possible?

OIDs are not guaranteed unique across databases.  To make this work
transparently, we'd have to change all internal APIs to use database
OID + object OID, not just object OID, as object identifiers.  That
would be horrendously invasive, not to mention probably costly in
performance (by doubling the width of cache keys, for instance).

But totally independently of implementation considerations, it's not
apparent to me that this would be a good idea.  As Frank mentioned,
Postgres has always treated databases as isolation boundaries.
Changing that would probably break a lot of peoples' installation
designs.

			regards, tom lane