Tyler Kellen <[EMAIL PROTECTED]> writes:
> I have a question about a full outer join returning duplicate rows.
Why do you think they are duplicate? The GROUP BY includes many columns
that you can't see directly in the output ...
Also, you did not show us the actual input data (where's the rows w
Tim Perdue <[EMAIL PROTECTED]> writes:
> SELECT *
> FROM (stats_http_downloads sh
> FULL OUTER JOIN stats_ftp_downloads sf USING (day,filerelease_id)) ;
> ERROR: FULL JOIN is only supported with mergejoinable join conditions
I've committed a fix for this problem. It'll appear in 7.1.1, or you
Tim Perdue <[EMAIL PROTECTED]> writes:
>> Is either of these tables actually a view?
> Hehe - no. I sent the \d of both tables at the bottom of that email.
\d isn't very helpful for these sorts of reports. How about pg_dump -s ?
regards, tom lane
--
On Fri, Apr 13, 2001 at 11:11:26AM -0400, Tom Lane wrote:
> Tim Perdue <[EMAIL PROTECTED]> writes:
> > Does it still work with those revised CREATE statements?
>
> Yes, works fine here.
>
> >> What version are you using?
>
> > That's RC3 and RC1.
>
> Curious. I'd expect this to happen for col
Tim Perdue <[EMAIL PROTECTED]> writes:
> Does it still work with those revised CREATE statements?
Yes, works fine here.
>> What version are you using?
> That's RC3 and RC1.
Curious. I'd expect this to happen for column datatypes whose '='
operator isn't marked mergejoinable, but plain old int
On Fri, Apr 13, 2001 at 03:02:32AM -0400, Tom Lane wrote:
> Tim Perdue <[EMAIL PROTECTED]> writes:
> > However, I keep getting this error, and I don't know what it means:
> > ERROR: FULL JOIN is only supported with mergejoinable join conditions
>
> Works for me:
>
> regression=# create table st
Tim Perdue <[EMAIL PROTECTED]> writes:
> However, I keep getting this error, and I don't know what it means:
> ERROR: FULL JOIN is only supported with mergejoinable join conditions
Works for me:
regression=# create table stats_http_downloads (day int, filerelease_id int);
CREATE
regression=# cr
