On 12/28/21 8:26 PM, Joel Jacobson wrote:
> On Mon, Dec 27, 2021, at 19:15, Tom Lane wrote:
>> NATURAL JOIN is widely regarded as a foot-gun that the SQL committee
>> should never have invented. Why would we want to create another one?
>>
>> (I suspect that making the constraint name optional would be problematic
>> for reasons of syntax ambiguity, anyway.)
>
> I agree. I remember this blog post from 2013 discussing the problems
> with both NATURAL but also the problems with USING:
> http://www.databasesoup.com/2013/08/fancy-sql-monday-on-vs-natural-join-vs.html
>
> Since my last email in this thread, I've learned KEY is unfortunately not a reserved keyword.
> This probably means the proposed "JOIN KEY" would be problematic, since a relation could be named KEY.
>
> Can with think of some other suitable reserved keyword?
I don't particularly like this whole idea anyway, but if we're going to
have it, I would suggest
JOIN ... USING KEY ...
since USING currently requires a parenthesized list, that shouldn't
create any ambiguity.
> How about JOIN WITH?
WITH is severely overloaded already.
--
Vik Fearing