Re: [HACKERS] Idea on how to simplify comparing two sets - Mailing list pgsql-hackers

From Tom Lane
Subject Re: [HACKERS] Idea on how to simplify comparing two sets
Date
Msg-id 5212.1486575022@sss.pgh.pa.us
Whole thread Raw
In response to Re: [HACKERS] Idea on how to simplify comparing two sets  (David Fetter <david@fetter.org>)
Responses Re: [HACKERS] Idea on how to simplify comparing two sets  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-hackers
David Fetter <david@fetter.org> writes:
> On Wed, Feb 08, 2017 at 11:22:56AM -0500, Tom Lane wrote:
>> Yes.  I think a new set-operation keyword would inevitably have to
>> be fully reserved --- UNION, INTERSECT, and EXCEPT all are --- which
>> means that you'd break every application that has used that word as
>> a table, column, or function name.

> I've long wanted a SYMMETRIC DIFFERENCE join type, that being the only
> elementary set operation not included in join types, but nobody at the
> SQL standards committee seems to have cared enough to help.

I wonder whether you could shoehorn it in with no new reserved word
by spelling it "EXCEPT SYMMETRIC", which could be justified by the
precedent of BETWEEN SYMMETRIC.  But not sure what to do with
duplicate rows (ie, if LHS has two occurrences of row X and RHS
has one occurrence, do you output X?)
        regards, tom lane



pgsql-hackers by date:

Previous
From: David Fetter
Date:
Subject: Re: [HACKERS] Idea on how to simplify comparing two sets
Next
From: "David G. Johnston"
Date:
Subject: Re: [HACKERS] Idea on how to simplify comparing two sets