Question

I am having trouble getting this complicated select to work for me.

Here is a graphic of my structure.
http://www.bandwise.com/files/tester_db.png

SQL file of the structure and sampe data.
http://www.bandwise.com/files/tester_db.sql

To put is simply, I have a table called “leases” and one called “apartments.” There is a one to many relationship between leases and apartments. Apartments identify a U.S. state and county in which they belong. Therefore, a given lease may have many apartments scattered across the country.

What I need to do is select ONLY U.S. states which contain one or more apartments and return a count of how many leases apply. Because some leases may have apartments in multiple states, a lease may need to be counted in multiple states.

Here is a test query that I’m working with.

SET @timenow = 1360855314;

SELECT l.lid, c.aid, c.leaseid, c.serialnum, c.state, c.county
FROM leases l, apartments c
WHERE c.leaseid = l.lid
    AND l.closebidding > @timenow AND l.status = 1;

I can see here that I have 6 leases in state #1 (Texas) and 2 leases in state #2 (Louisiana). How do I go about querying from the states table outward to get the number of leases based on where their apartments are located?

The return should look like this.

---------------------------
|  state      |   total   |
---------------------------
|  Texas      |     6     |
---------------------------
|  Louisiana  |     3     |
---------------------------

Answer 1

Use a count(distinct l.lid):

SELECT s.state_name,count(distinct l.lid) 
FROM 
  leases l
  JOIN apartments c ON c.leaseid = l.lid
  JOIN states s ON s.state_id = c.state
WHERE
  l.closebidding > @timenow AND l.status = 1 
GROUP BY s.state_name;

SQL Fiddle