Re: Add a ROWCOUNT to the output of a select.

On Wed, May 14, 2008 at 5:24 AM, Gavin 'Beau' Baumanis <
gavinb(at)eclinic(dot)com(dot)au> wrote:

> Hi Everyone,
>
> After spending some time searching through our good friend Mr. Google and
> the mailing list I found a post that provided me with a query that does just
> what I need.
>
> However, the query takes FOREVER and although this is stated in the
> original mail I thought I would ask if there was any advice that you might
> be able to provide to speed things up a little.
>
> And while the query does return over 27,000 rows in my case, I didn't
> expect it to take 9 minutes and 11 seconds!
>
> Please find the query below - and of course - thanks in advance for any
> assistance you might be able to provide me with!
>
> select
> (select
> count(*)
> from
> myTable as myCount
> where
> myCount.contactdate <= myTable.contactdate
> ) as rownum,
> contactdate
> from
> myTable
> where
> contactdate > '2007-06-30 23:59:59'
> order by
> contactdate;
>
> <http://www.postgresql.org/mailpref/pgsql-sql>

Posting EXPLAIN command's output would have helped here.

This sub-query in the SELECT list is a correlated sub-query, so it will be
executed for each row that passes the oouter query's WHERE clause. If you
don't have it already, I'd suggest creating an index on the 'contactdate'
column; that should help speed up the query. In absence of such an index,
the planner will choose Sequential Scan, which is very
expensive/time-consuming.

Best regards,
--
gurjeet[(dot)singh](at)EnterpriseDB(dot)com
singh(dot)gurjeet(at){ gmail | hotmail | indiatimes | yahoo }.com

EnterpriseDB http://www.enterprisedb.com

Mail sent from my BlackLaptop device