"SELECT *" vs hidden columns and logical column order

From: Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com>
To: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: "SELECT *" vs hidden columns and logical column order
Date: 2017-06-29 03:52:10
Message-ID: CAEepm=3ZHh=p0nEEnVbs1Dig_UShPzHUcMNAqvDQUgYgcDo-pA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi hackers,

I am aware of at three potential projects that would change the
meaning of "SELECT *":

1. Incremental MATERIALIZED VIEW maintenance probably needs to be
able to use a hidden counter column which you can ask for by name but
will otherwise not show up to users:

https://www.postgresql.org/message-id/1371480075.55528.YahooMailNeo@web162901.mail.bf1.yahoo.com

2. SQL:2011 temporal tables track system time and/or valid time with
columns that users create and then declare to be temporal control
columns. I don't think they show up unless you name them directly (I
didn't check the standard but I noticed that it's that way in another
product), so I guess that's basically the same as (1).

3. Logical column order aka ALTER COLUMN POSITION, a recurring topic
on pgsql-hackers for which patches have been written but nothing has
so far managed to stick:

https://www.postgresql.org/message-id/flat/20141209174146.GP1768%40alvh.no-ip.org

Suppose someone wanted to chip away at a small piece of incremental
matviews by inventing a way to declare 'hidden' columns: is there
really a dependency here as implied in the 2013 email above? Is
anyone planning to revive logical column order?

--
Thomas Munro
http://www.enterprisedb.com

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Craig Ringer 2017-06-29 04:23:00 Re: protocol version negotiation (Re: Libpq PGRES_COPY_BOTH - version compatibility)
Previous Message Mengxing Liu 2017-06-29 03:36:34 [GSOC][weekly report 4] Eliminate O(N^2) scaling from rw-conflict tracking in serializable transactions