Re: Plans for index names unique to a table?

From: Christopher Browne <cbbrowne(at)cbbrowne(dot)com>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Plans for index names unique to a table?
Date: 2003-05-10 17:44:42
Message-ID: 20030510174442.275503E874@cbbrowne.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Jim C. Nasby wrote:
> FWIW, DB2 and (I think) Oracle both use a global namespace for
> indexes. Sybase and (I think) MSSQL don't. I really don't understand
> why you'd want indexes in a global namespace and wish IBM and Oracle
> would change their products.

There are at least four compelling disadvantages to implementing a
"table" namespace for indices:

a) It is a change from current functionality, and would break things
that work now;

b) It introduces a new namespace, and the need to manage it.

c) It is incompatible with several other databases' treatment of
indices, and probably the SQL standards, too.

d) It introduces the need to change syntax on indices, requiring
something like
DROP INDEX "index_name" ON TABLE "some_table";
which parallels the handling of RI triggers.

They may not all be individually compelling, but they add up...
--
output = ("cbbrowne" "@acm.org")
http://www3.sympatico.ca/cbbrowne/lisp.html
"More computing sins are committed in the name of efficiency (without
necessarily achieving it) than for any other single reason - including
blind stupidity." -- W.A. Wulf

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Ron Johnson 2003-05-10 18:31:30 Re: realtime data inserts
Previous Message Jim C. Nasby 2003-05-10 17:12:23 Re: Plans for index names unique to a table?