Date: Tue, 27 Jan 2004 23:36:55 -0700
From: j knight <enabled@myrealbox.com>
Subject: Aggregate function error in 7.4
To: pgsql-sql@postgresql.org
Message-id: <40175887.2080108@myrealbox.com>
MIME-version: 1.0
Content-type: text/plain; charset=us-ascii; format=flowed
Content-transfer-encoding: 7BIT
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6)
	Gecko/20040113

Hi all.

I'm trying to migrate from 7.3.4 to 7.4.1. I've performed a pg_dumpall 
using the pg_dumpall from 7.4.1. Once 7.4 is running and I try to import 
the dump, I get the following error:

psql:pgsql.dump:301: ERROR:  column "r.day" must appear in the GROUP BY 
clause or be used in an aggregate function


The SQL it's choking on is:

CREATE VIEW maillog_day_tally AS
SELECT
     day,
     COALESCE(sum(r.tally), 0) AS received,
     COALESCE(sum(s.tally), 0) AS sent
FROM
     maillog_recv r FULL JOIN maillog_sent s USING (day, address)
GROUP BY day;


The tables:

CREATE TABLE maillog_sent (
     day date NOT NULL,
     tally integer NOT NULL,
     address text NOT NULL,
     CONSTRAINT maillog_sent_pk PRIMARY KEY (day, address)
);

CREATE TABLE maillog_recv (
     day date NOT NULL,
     tally integer NOT NULL,
     address text NOT NULL,
     CONSTRAINT maillog_recv_pk PRIMARY KEY (day, address)
);


Of course if I change the query to read "SELECT r.day ..." it will 
parse, however, I need to be able to select the day column from the 
result of the JOIN, not just from one of the two tables.

I'm at a loss to explain why 7.4 is treating that SELECT query 
differently than previous versions. I can't see anything in the 7.4 
release notes that would help explain this. Can anyone provide any insight?


maillog=> select version();
version
---------------------------------------------------------------------
PostgreSQL 7.4.1 on i386-unknown-freebsd4.7, compiled by GCC 2.95.4


.joel