On Fri, Jul 30, 2010 at 11:50 AM, Merlin Moncure <mmoncure(at)gmail(dot)com> wrote:
> On Fri, Jul 30, 2010 at 10:34 AM, Derrick Rice <derrick(dot)rice(at)gmail(dot)com> wrote:
>> Is it possible to use the ARRAY(select ...) syntax as a substitute for
>> array_agg on versions of postgresql that don't have it? (8.2) It works
>> simply enough when only selecting a single column, but if I need to group by
>> some other column, I'm not clear how I'd go about doing that.
>> For example, write the following in ARRAY(select...) form.
> yup...we've had array() for ages (I think -- see below).
> now, you've always been able to do array aggregation in userland --
> it's been in the docs as example since I can remember (see here:
> The old school array_accum however is much slower than the newer
> array_agg. (which in turn is just a tiny bit slower than array()
> discounting grouping effects). So you _definitely_ want to use
> array() if you don't require aggregate grouping features in older
> postgres versions.
> hm. I looked for the documentation for array(select...) to figure out
> exactly when it was introduced, couldn't find it. Anyone know
> if/where this is documented?
I found it -- array() syntax is documented in array_constructor
portion of syntax.sgml. It's been in postgres at least since 7.4,
which is as far back as I checked.
In response to
pgsql-general by date
|Next:||From: Joshua Tolley||Date: 2010-08-01 04:16:57|
|Subject: Re: PostgreSQL and distributed transactions|
|Previous:||From: Devrim GÜNDÜZ||Date: 2010-07-31 10:03:46|
|Subject: Re: Centos 5.5 and Postgres 9.0.beta3 install errors|