Yes, I like Chris's suggestion.

For example, in PostgreSQL, I can do the following:

postgres=# select * from ddd;
call_date_time | in_trunk
----------------------------+----------
2004-02-10 17:10:47.95334 | 25
2004-02-10 17:10:53.111503 | 32
2004-02-10 17:10:57.444058 | 100

postgres=# select * from dddd order by 1 DESC;
call_date_time1 | call_date_time2
----------------------------+-------------------------------
2004-02-10 17:10:47.95334 | 2004-02-10 17:10:47.95334-08
2004-02-07 23:27:30.365545 | 2004-02-07 23:27:30.365545-08
(2 rows)

postgres=# select * from ddd where call_date_time in (select call_date_time1
from dddd order by 1 DESC limit 1);
call_date_time | in_trunk
---------------------------+----------
2004-02-10 17:10:47.95334 | 25
(1 row)

Regards,
Dorn.

----- Original Message -----
From: "Chris Hall" <Chris.Hall@orbisuk.com>
To: <informix-list@iiug.org>
Sent: Friday, February 13, 2004 07:30 AM
Subject: Re: Whatcha' wanta have?????


> jleffler@us.ibm.com (Jonathan Leffler) wrote in message
news:<eab66e61.0402121022.7f9e03d2@posting.google. com>...
> > "Dirk Moolman" <DirkM@mxgroup.co.za> wrote:
> > > This makes me think of another simple thing:
> > >
> > > Increasing the capability of: select first n * from tablename
> > >
> > > ..... the capability of selecting subsets of data without too much
trouble
> >
> > Can you explain what you're thinking of? What do you want that Dorn
> > Bhechsonggram's suggestion does not provide? Or where do you want to
> > use the notation that you can't in PostgreSQL and MySQL?
> >
> > Thanks for the extra information.
> >
> > And thanks for the valuable ideas you're all submitting. No promises,
> > of course, but we're definitely taking notes.
> >
> > -=JL=-
>
> being able to use "first N" in any select, e.g.
>
> select * from table_1 where key in (select first 10 key from table_2
> order by something);
>
> select first N * from table_1 into temp table_temp;
>
> etc...
>


sending to informix-list


sending to informix-list