Here it is:

select @row := @row + 1 as row, uid, comment_count from (SELECT c.uid, COUNT(c.cid) AS comment_count FROM comments c WHERE c.uid <> 0 GROUP BY c.uid ORDER BY comment_count DESC) as com_count, (SELECT @row := 0) r

Note: I've changed comments.mail to comments.uid, since my testing db has mail fields empty.

But, I am warning you that this query might brake if you put it through db_query_rewrite.

On Fri, 2010-10-29 at 05:35 -0700, nan wich wrote:

Thanks, but that doesn't do what I want. What I need is the row number that results from this query (simplified for example).

SELECT c.mail, COUNT(c.cid) AS comment_count FROM comments c WHERE c.mail <> '' GROUP BY c.mail ORDER BY comment_count DESC

And remember that pager_query is going to add LIMIT xxx, 10.

Nancy

Injustice anywhere is a threat to justice everywhere. -- Dr. Martin L. King, Jr.

From: Nikola Kotur <kotnick@gmail.com>
To: development@drupal.org
Sent: Fri, October 29, 2010 5:19:17 AM
Subject: Re: [development] Ranking in a query

On Thu, 2010-10-28 at 06:32 -0700, nan wich wrote:

I have a query that counts comments and all is well with it. But I'd like add a ranking column to the display. The problem is that it uses tablesort, so the ranking has to be in the query because I don't know which column will be used to sort the display. Oh, it is also paged so I can't count on the first row being #1 and so on. Has anyone got an idea on how to add a ranking column to the query? (MySql only.)

Nancy, here's a query that returns a nid, and it's row position:

SELECT @row := @row + 1 as row, n.nid FROM node n, (SELECT @row := 0) r;

You can start from there.