Re: [sqlsmith] Failed to generate plan on lateral subqueries

From: Andreas Seltenreich <seltenreich(at)gmx(dot)de>
To: Peter Geoghegan <pg(at)heroku(dot)com>
Cc: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Subject: Re: [sqlsmith] Failed to generate plan on lateral subqueries
Date: 2015-12-11 16:47:26
Message-ID: 87bn9xhqzl.fsf@elite.ansel.ydns.eu
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Peter Geoghegan writes:

> On Sun, Dec 6, 2015 at 9:52 AM, Andreas Seltenreich <seltenreich(at)gmx(dot)de> wrote:
>> I've added new grammar rules to sqlsmith and improved some older ones.
>
> Could you possibly teach sqlsmith about INSERT ... ON CONFLICT DO
> UPDATE/IGNORE? I think that that could be very helpful, especially if
> it could be done in advance of any stable release of 9.5.

In summary, it can't be added ad-hoc, but might still happen in advance
of the release of 9.5.

Adding upsert needs significiant effort because historically,
non-boolean value expression productions yield a random type. This is
not a problem for generating queries, but it is for inserts. Also,
sqlsmith can at the moment only generate sensible value expressions from
column references. Generating a proper upsert would require supporting
type-constraining of productions as well as adding productions for
pulling values out of thin air (e.g., generating atomic value subselects
or calling argumentless functions).

regards,
Andreas

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Masahiko Sawada 2015-12-11 16:56:50 Re: Tab-comletion for RLS
Previous Message Robert Haas 2015-12-11 16:20:21 Re: Rework the way multixact truncations work