Re: [PATCH] Add support for INSERT ... SET syntax - Mailing list pgsql-hackers

From solai v
Subject Re: [PATCH] Add support for INSERT ... SET syntax
Date
Msg-id CAF0whucy3wpEYqtVEX=eKi9rd5uR-=hQMkaCmjYwxQ-rOmheVw@mail.gmail.com
Whole thread
In response to Re: [PATCH] Add support for INSERT ... SET syntax  (Álvaro Herrera <alvherre@kurilemu.de>)
List pgsql-hackers
Hi Suraj,

I tested v3 of the patch on current master.
The patch applied and built cleanly on my setup. Before applying it,
statements using the proposed INSERT...SET syntax resulted in a syntax
error near SET, which is expected since PostgreSQL does not currently
support this syntax. After applying the patch, the same statements
worked as expected and rows were inserted successfully.
I went through the functionality added by the patch and tested a few
different scenarios, including basic INSERT...SET usage, DEFAULT
values, expressions/functions, RETURNING, ON CONFLICT DO UPDATE,
multi-row inserts, different column sets in multi-row inserts, and
subqueries in assignments.
All of them behaved as expected in my testing.
I also verified the query that Ajay previously reported as causing a
backend crash:
INSERT INTO emp_test SET (empno,ename)=(SELECT 1,'aa');
With v3, I was not able to reproduce the crash. The query returned a
normal syntax error, and the server continued to run normally
afterward.
In addition, I ran the regression test suite and all tests passed successfully.
Overall, the patch worked well in my testing and I did not notice any
regressions.


Regards,
solai



pgsql-hackers by date:

Previous
From: Bertrand Drouvot
Date:
Subject: Re: [PATCH] doc: clarify pg_stat_lock.fastpath_exceeded scope
Next
From: vignesh C
Date:
Subject: Re: Proposal: Conflict log history table for Logical Replication