Add domain to TODO.detail

This commit is contained in:
Bruce Momjian 2001-06-04 20:13:01 +00:00
parent 8b74f70d68
commit cadf7ee34e

142
doc/TODO.detail/domain Normal file
View File

@ -0,0 +1,142 @@
From pgsql-hackers-owner+M8916@postgresql.org Thu May 17 11:51:45 2001
Return-path: <pgsql-hackers-owner+M8916@postgresql.org>
Received: from postgresql.org (webmail.postgresql.org [216.126.85.28])
by candle.pha.pa.us (8.10.1/8.10.1) with ESMTP id f4HFpid02613
for <pgman@candle.pha.pa.us>; Thu, 17 May 2001 11:51:44 -0400 (EDT)
Received: from postgresql.org.org (webmail.postgresql.org [216.126.85.28])
by postgresql.org (8.11.3/8.11.1) with SMTP id f4HFoiA36759;
Thu, 17 May 2001 11:50:44 -0400 (EDT)
(envelope-from pgsql-hackers-owner+M8916@postgresql.org)
Received: from fred.plzen-city.cz (gate.plzen-city.cz [194.212.191.10])
by postgresql.org (8.11.3/8.11.1) with ESMTP id f4HFMLA24019
for <pgsql-hackers@postgreSQL.org>; Thu, 17 May 2001 11:22:21 -0400 (EDT)
(envelope-from horak@sit.plzen-city.cz)
Received: from exchange.mmp.plzen-city.cz ([192.168.4.42])
by fred.plzen-city.cz (8.12.0.Beta7/8.12.0.Beta7/Wim) with ESMTP id f4HFMGoQ007660
for <pgsql-hackers@postgreSQL.org>; Thu, 17 May 2001 17:22:16 +0200
Subject: [HACKERS] possible DOMAIN implementation
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-2"
Content-Class: urn:content-classes:message
X-MimeOLE: Produced By Microsoft Exchange V6.0.4418.65
Date: Thu, 17 May 2001 17:22:05 +0200
Message-ID: <9C8918CC8CDAAC4AA79D4594A480648901E68C2B@EXCHANGE.mmp.plzen-city.cz>
Thread-Topic: possible DOMAIN implementation
Thread-Index: AcDe5WK0+5JgskVMEdWdCQAQS2oI5Q==
From: =?iso-8859-2?Q?Hor=E1k_Daniel?= <horak@sit.plzen-city.cz>
To: <pgsql-hackers@postgresql.org>
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by postgresql.org id f4HFMMA24022
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org
Status: OR
Hello,
I have spend some thinking about implementation of DOMAIN capability.
Here are my ideas.
What is a domain? It is an alias for a type with size, constraints and
default values. It is like one column of a table. And this is the main
idea of my "implementation". It should be possible to implement it using
existing system tables.
New rules for grammar can be easily created from already existing pieces
(column definition of a table).
How to store information about a domain in system tables?
When a new domain is created it will:
- put a record into pg_type with typnam = domain name, new code for
typtype = 'd' and typrelid = oid of a new record in pg_class (next line)
- put a record into pg_class to create a fictional table with a new
relkind ('d'?), relnatts = 1, relname can be system generated
(pg_d_<domainname>)
- put a records into pg_attribute and pg_attrdef with "column
(attribute) definition" - real type, size, default value etc., owner
will the fictional table from the previous step
Then it will be required to modify functions that works with types. When
typtype of a retrieved type is 'd' then it will perform lookups into
pg_class, pg_attribute and pg_attrdef to find the real definition of the
domain. These additional lookups will also create a performace penalty
of using domains. But every feature has its costs. I know this paragraph
about the real implementation is very short, but I think there are
people that know the "type mechanism" better then I know. And can easier
tell if it is possible to go this way.
I hope you understand my explanation. It is also possible that I don't
know some aspects of the backend code that makes my idea wrong.
Dan
----------------------------------------------
Ing. Daniel Horak
network and system administrator
e-mail: horak@sit.plzen-city.cz
privat e-mail: dan.horak@email.cz ICQ:36448176
----------------------------------------------
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
From pgsql-hackers-owner+M8925@postgresql.org Thu May 17 14:04:29 2001
Return-path: <pgsql-hackers-owner+M8925@postgresql.org>
Received: from postgresql.org (webmail.postgresql.org [216.126.85.28])
by candle.pha.pa.us (8.10.1/8.10.1) with ESMTP id f4HI4Sd12032
for <pgman@candle.pha.pa.us>; Thu, 17 May 2001 14:04:28 -0400 (EDT)
Received: from postgresql.org.org (webmail.postgresql.org [216.126.85.28])
by postgresql.org (8.11.3/8.11.1) with SMTP id f4HI3hA18949;
Thu, 17 May 2001 14:03:43 -0400 (EDT)
(envelope-from pgsql-hackers-owner+M8925@postgresql.org)
Received: from sss.pgh.pa.us (sss.pgh.pa.us [216.151.103.158])
by postgresql.org (8.11.3/8.11.1) with ESMTP id f4HHx4A16857
for <pgsql-hackers@postgresql.org>; Thu, 17 May 2001 13:59:04 -0400 (EDT)
(envelope-from tgl@sss.pgh.pa.us)
Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1])
by sss.pgh.pa.us (8.11.3/8.11.3) with ESMTP id f4HHwjR12171;
Thu, 17 May 2001 13:58:45 -0400 (EDT)
To: =?iso-8859-2?Q?Hor=E1k_Daniel?= <horak@sit.plzen-city.cz>
cc: pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] possible DOMAIN implementation
In-Reply-To: <9C8918CC8CDAAC4AA79D4594A480648901E68C2B@EXCHANGE.mmp.plzen-city.cz>
References: <9C8918CC8CDAAC4AA79D4594A480648901E68C2B@EXCHANGE.mmp.plzen-city.cz>
Comments: In-reply-to =?iso-8859-2?Q?Hor=E1k_Daniel?= <horak@sit.plzen-city.cz>
message dated "Thu, 17 May 2001 17:22:05 +0200"
Date: Thu, 17 May 2001 13:58:45 -0400
Message-ID: <12167.990122325@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org
Status: OR
=?iso-8859-2?Q?Hor=E1k_Daniel?= <horak@sit.plzen-city.cz> writes:
> When a new domain is created it will:
> - put a record into pg_type with typnam = domain name, new code for
> typtype = 'd' and typrelid = oid of a new record in pg_class (next line)
> - put a record into pg_class to create a fictional table with a new
> relkind ('d'?), relnatts = 1, relname can be system generated
> (pg_d_<domainname>)
Ugh. Don't overload pg_class with things that are not tables. I see no
reason that either pg_class or pg_attribute should be involved in the
definition of a domain. Make new system tables if you need to, but
don't confuse the semantics of critical tables.
> - put a records into pg_attribute and pg_attrdef with "column
> (attribute) definition" - real type, size, default value etc., owner
> will the fictional table from the previous step
> Then it will be required to modify functions that works with types. When
> typtype of a retrieved type is 'd' then it will perform lookups into
> pg_class, pg_attribute and pg_attrdef to find the real definition of the
> domain. These additional lookups will also create a performace penalty
> of using domains.
Why shouldn't this info be directly available from the pg_type row?
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org