Date: Wed, 2 Dec 2020 10:53:51 -0600
From: Justin Pryzby <pryzby@telsasoft.com>
To: Luc Vlaming <luc@swarm64.com>
Cc: Simon Riggs <simon@2ndquadrant.com>, Andres Freund <andres@anarazel.de>,
	pgsql-hackers@lists.postgresql.org,
	Heikki Linnakangas <hlinnaka@iki.fi>
Subject: Re: should INSERT SELECT use a BulkInsertState?
Message-ID: <20201202165351.GG24052@telsasoft.com>
References: <20200604173047.dirdp6zln7ogj54k@alap3.anarazel.de>
 <20200713015700.GA23581@telsasoft.com>
 <20200919133215.GB30557@telsasoft.com>
 <20201016210511.GC9241@telsasoft.com>
 <CANP8+jKmvtaq8a1YkKsfXpWOc3keN2k7K84Su+y--WZO6D_jFQ@mail.gmail.com>
 <20201030045138.GA3080@telsasoft.com>
 <bd0dc980-c4c9-29c9-8cb6-8d7787ae9fef@swarm64.com>
 <20201102184551.GJ22691@telsasoft.com>
 <20201124020020.GK24052@telsasoft.com>
 <20201130011146.GS24052@telsasoft.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20201130011146.GS24052@telsasoft.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Precedence: bulk

One loose end in this patch is how to check for volatile default expressions.

copyfrom.c is a utility statement, so it can look at the parser's column list:
COPY table(c1,c2)...

However, for INSERT, in nodeModifyTable.c, it looks like parsing, rewriting,
and planning are done, at which point I don't know if there's a good way to
find that.  The default expressions will have been rewritten into the planned
statement.

We need the list of columns whose default is volatile, excluding columns for
which a non-default value is specified.

INSERT INTO table (c1,c2) VALUES (1,default);

We'd want the list of any column in the table with a volatile default,
excluding columns c1, but not-excluding explicit default columns c2 or any
implicit default columns (c3, etc).

Any idea ?

-- 
Justin