Question

OK, a slight variation on an earlier theme. Using the same basic idea, I want to get independent counts of the fields, then I want them grouped by a higher order breakdown.

I expanded the example by David to include a higher order column:

district_id, product_id, service_id

dist    proj    serv
1   1   1
1   1   2
1   1   2
1   1   3
1   1   3
1   1   4
1   2   2
1   2   4
1   2   4
1   2   5
1   2   5
2   1   1
2   2   1
2   1   6
2   2   6
2   3   6

To get a result on the total, I used a simple query with two sub-queries.

select 
(select count(Distinct project_id) from GroupAndCountTest) AS "projects",
(select count(Distinct service_id) from GroupAndCountTest) as "services";

projects  services
       3         6

The challenge was to get this grouped within the district_id. What I wanted was:

district_id   projects  services
          1          2         5
          2          3         6 

I ended up using similar sub-queries, but the only way I was able to combine them (other than using a stored function) was to re-run the sub-queries for every district. (Not a big problem here, but in my application the sub-queries use multiple tables with a substantial number of “districts” so the two sub-queries are run again for each “district” which will become increasingly ineffecient.

This query works, but I would love to see something more effecient.

select t1.district_id, p1.projects, s1.services
  from GroupAndCountTest as t1
  join (select district_id, count(Distinct project_id) as projects 
    from GroupAndCountTest
    group by district_id) AS p1
    on p1.district_id=t1.district_id
  join (select district_id, count(Distinct service_id) as services 
    from GroupAndCountTest
    group by district_id) as s1
    on s1.district_id=t1.district_id
  group by t1.district_id;

Thanks.

PS: If you want to experiment, you can create the table with:

CREATE TABLE `GroupAndCountTest` (
  `district_id` int(5) DEFAULT NULL,
  `project_id` int(5) DEFAULT NULL,
  `service_id` int(5) DEFAULT NULL
) ENGINE=MyISAM DEFAULT CHARSET=latin1;
insert  into `GroupAndCountTest`(`district_id`,`project_id`,`service_id`) 
  values (1,1,1),(1,1,2),(1,1,2),(1,1,3),(1,1,3),(1,1,4),(1,2,2),(1,2,4),
  (1,2,4),(1,2,5),(1,2,5),(2,1,1),(2,2,1),(2,1,6),(2,2,6),(2,3,6);

Answer 1

select district_id, 
count(distinct(product_id)) projects, 
count(distinct(service_id)) services 
from MyTable group by district_id;

where MyTable contains district_id, product_id, service_id columns