Re: joined tables with USING and GROUPBY on the USING() column
От | John R Pierce |
---|---|
Тема | Re: joined tables with USING and GROUPBY on the USING() column |
Дата | |
Msg-id | e4101f69-1cef-e961-7ffb-d5ba1c6cbec0@hogranch.com обсуждение исходный текст |
Ответ на | joined tables with USING and GROUPBY on the USING() column (Geoff Winkless <pgsqladmin@geoff.dj>) |
Список | pgsql-general |
On 9/7/2016 12:53 PM, Geoff Winkless wrote: > I'll start by saying that I'm sure I'm missing something obvious... > > I have a query that is working fine on all my servers except one. The > only obvious difference is that the failing one is running 9.5.3, > while most are running 9.5.4, but since the query works on a 9.5.1 box > I also have I can't imagine that's the cause. > > I can simplify the query (and still retain the error) down to: > > SELECT results.token FROM results INNER JOIN tokens USING(token) GROUP BY token; > > ERROR: column "results.token" must appear in the GROUP BY clause or be > used in an aggregate function > > All well and good, and I'm happy enough to change the query's GROUP BY > to include the table name, but it's confusing me how it works OK on > all servers except one. > > Is there some configuration option I'm missing? the only thing I can think of is that 'group by token' is ambiguous, if that was group by results.token, it would be specific. yes, I know, your join conditions using(token) are equivalent to saying on results.token=tokens.token, so either token should be the same thing... -- john r pierce, recycling bits in santa cruz
В списке pgsql-general по дате отправления: