Re: Support UPDATE table SET(*)=...

Peter Geoghegan <pg(at)heroku(dot)com> writes:
> On Tue, Apr 7, 2015 at 11:19 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> If we were to go in this direction, it would be nice to at the same time
>> add a similar whole-record syntax for INSERT. I'm not sure exactly what
>> that should look like though. Also, again, we ought to be paying
>> attention to how this would match up with UPSERT syntax.

> I expressed concern about allowing this for UPSERT [1].

Yeah, your analogy to "SELECT *" being considered dangerous in production
is not without merit. However, to the extent that the syntax is used to
assign from a composite variable of the same (or compatible) data type,
it seems like it would be safe enough. IOW, I think that analogy holds
for the syntax implemented by the current patch, but not what I suggested
in my followup.

regards, tom lane