Re: 7.3.3 behaving differently on OS X 10.2.6 and FreeBSD

I have a suspicion that the version might be different. I have the same
symptom here on two different RH 7.3 boxes one running 7.3.2 and the
other running 7.3.3
It would appear 7.3.2 is more strict about the naming of the GROUP BY
fields.

David Olbersen wrote:

>Culley,
>
>
>
>>But on my production machine postgresql complained about the order by
>>clause-- it wanted the table alias to be on last_name.
>>
>>
>
>I believe this is because you used "u.last_name" earlier in the statement, and the ORDER BY clause doesn't know that's what you mean.
>
>That's a guess that doesn't really explain why it'd work under one OS and not under another. Are the two versions of Postgres configured the same?
>
>--------------------------
>David Olbersen
>iGuard Engineer
>St. Bernard Software
>
>
>
>
>>-----Original Message-----
>>From: culley harrelson [mailto:culley(at)fastmail(dot)fm]
>>Sent: Friday, August 08, 2003 10:48 AM
>>To: pgsql-general(at)postgresql(dot)org
>>Subject: [GENERAL] 7.3.3 behaving differently on OS X 10.2.6
>>and FreeBSD
>>4.8-STABLE
>>
>>
>>I don't know if this is a postgresql bug or a problem with my
>>architecture but I thought I would post here about a strange
>>bug I just
>>came across in my application.
>>
>>I use OS X 10.2.6 as my development machine and FreeBSD 4.8 for my
>>production machines. All systems are running postgresql
>>7.3.3. I just
>>published some code to production and when testing the production
>>results it blew up with a sql parsing error. The following
>>sql worked
>>fine on my OS X development machine:
>>
>>select u.user_id, u.first_name, u.last_name, u.email_address, w.w9,
>>pm.description as payment_method, count(s.user_id) as documents,
>>sum(s.payment_amount) as amt_sum from ht_user u inner join
>>writer w on
>>u.user_id = w.user_id inner join payment_method pm on
>>w.payment_method_id = pm.payment_method_id left join submission s on
>>u.user_id = s.user_id group by u.user_id, u.first_name, u.last_name,
>>u.email_address, w.w9, pm.description order by lower(last_name) asc
>>
>>But on my production machine postgresql complained about the order by
>>clause-- it wanted the table alias to be on last_name.
>>
>>culley
>>
>>
>>
>>---------------------------(end of
>>broadcast)---------------------------
>>TIP 7: don't forget to increase your free space map settings
>>
>>
>>
>
>---------------------------(end of broadcast)---------------------------
>TIP 8: explain analyze is your friend
>
>