Question

I’d like to select a new column named sliced (value can be 1/0 or true/false it doesn’t matter) if area of the current row equals MAX(SUM(c.area)), that is flag the row with highest aggregate value:

SELECT p.name AS name, SUM(c.area) AS area
FROM City AS c
   INNER JOIN Province AS p ON c.province_id = p.id
      INNER JOIN Region AS r ON p.region_id = r.id
WHERE r.id = ?
GROUP BY p.id
ORDER BY p.name ASC

I’ve tried adding to the selection area = MAX(area) AS sliced or even area = SUM(MAX(c.area)) AS sliced but i’m getting a syntax error. I’ve to admit i’m not so good in SQL. Thank you.

Answer 1

Here’s a way to do it with just one group by:

set @row := 0;
select name, area, sliced
from (
    select name, area, (@row := @row + 1) = 1 as sliced
    from (
        SELECT p.name, SUM(c.area) AS area
        FROM City AS c
        INNER JOIN Province AS p ON c.province_id = p.id
        INNER JOIN Region AS r ON p.region_id = r.id
        WHERE r.id = ?
        GROUP BY 1
        ORDER BY 2 desc) t1
    ) t2
order by 1;

The inner query (t1) does the group by and orders by total area largest first.
The next query (t2) gives the first row a value of true for column sliced, all other rows false.
The outer query orders the rows in the way you want – by name.

Since there’s only one table scan and group by, this should be very efficient.