Hi,
Tom Lane schrieb:
> [ experiments... ] This works reliably in 7.4 and up. Before that,
> the optimizer didn't make the connection between the sort ordering of
> the inner query and that needed by the outer, so it would repeat the
> sort step using only key1 and very possibly destroy the key2
Richard Huxton schrieb:
> Search the mailing-list archives for "custom aggregate concat" and
> you'll quickly find an example of how to write your own custom aggregate
> (like SUM()).
>
> Warning - I don't think you can guarantee the order of elements in the
> aggregated sectors.
Thank you ver
Richard Huxton writes:
> Search the mailing-list archives for "custom aggregate concat" and
> you'll quickly find an example of how to write your own custom aggregate
> (like SUM()).
> Warning - I don't think you can guarantee the order of elements in the
> aggregated sectors.
In recent PG ve
Kai Hessing wrote:
The normal clause would look like:
SELECT c.companyname, s.sectorname FROM company c, sector s,
company_sector cs WHERE cs.cid = c.cid AND cs.sid = s.sid ORDER BY
c.companyname;
c.companyname | ??? (sectors)
---+
company1 | secto
Hi Folks,
I hope to have found the right group for my question. I have difficult
sql-task. I try to describe it simple:
We have a table 'company' with a cid and a table 'sector' with a sid.
They are connected m:n via a third table 'company_sector' which contians
csid, cid and sid.
The normal cl
