MySQL AVG and Grouping - mysql

Question

I'm struggling with a MySQL statement and was hoping for some guidance as I am close, but not quite there. I have a database that contains a table of property addresses and of property rental listings. The property addresses are related to a table or regions, which is related to a table of districts, which is then related to a table of suburbs.
I am trying to create a result which gives me the average rent in each suburb per month and by the number of bedrooms.
For example:
District Suburb Month YEAR YMD Bedrooms DataAverage
Nelson The Brook 01 2012 2012-01-01 00:00 1 190
Nelson The Brook 01 2012 2012-01-01 00:00 2 274
Nelson The Brook 01 2012 2012-01-01 00:00 3 341
Which I can then convert into a table as follows:
Average Rent
Beds by Suburb Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12
The Brook
1 $150 $245 $160 $285 $135 $370 $350
2 $330 $340 $380 $310 $335 $345 $355
3 $350 $380 $310 $395 $380 $350 $350
Inner City
1 $160 $245 $260 $285 $295 $300 $350
2 $360 $440 $480 $410 $535 $545 $555
3 $370 $480 $510 $595 $480 $450 $550
My Current SQL query is this:
SELECT d.name as District, s.name AS Suburb,
FROM_UNIXTIME(l.StartDate,'%m') AS Month,
FROM_UNIXTIME(l.StartDate,'%Y') AS YEAR,
FROM_UNIXTIME(l.StartDate, '%Y-%m-01 00:00') AS YMD,
p.Bedrooms,
REPLACE(FORMAT(AVG(l.RentPerWeek),0),',','') AS DataAverage
FROM properties p
LEFT JOIN listings l on l.property_id=p.id
LEFT JOIN regions r on p.region_id=r.id
LEFT JOIN districts d on d.region_id=r.id
LEFT JOIN suburbs s on s.district_id=d.id
WHERE FROM_UNIXTIME(l.StartDate) BETWEEN DATE(NOW()) - INTERVAL (DAY(NOW()) - 1) DAY - INTERVAL 11 MONTH AND NOW()
GROUP BY District, Suburb, Year, Month, Bedrooms
ORDER BY District, Suburb ASC, YMD ASC, Bedrooms ASC
Unfortunately what I am getting is the same result for each and every suburb. I think I may need to create a subquery SQL statement to get this to work properly, but I'm not entirely sure.
So I am getting something like this:
District Suburb Month YEAR YMD Bedrooms DataAverage
Nelson The Brook 01 2012 2012-01-01 00:00 1 190
Nelson The Brook 01 2012 2012-01-01 00:00 2 330
Nelson The Brook 01 2012 2012-01-01 00:00 3 350
Nelson The Brook 02 2012 2012-02-01 00:00 1 245
Nelson The Brook 02 2012 2012-02-01 00:00 2 340
Nelson The Brook 02 2012 2012-02-01 00:00 3 380
...
Nelson Inner City 01 2012 2012-01-01 00:00 1 190
Nelson Inner City 01 2012 2012-01-01 00:00 2 330
Nelson Inner City 01 2012 2012-01-01 00:00 3 350
Nelson Inner City 02 2012 2012-02-01 00:00 1 245
Nelson Inner City 02 2012 2012-02-01 00:00 2 340
Nelson Inner City 02 2012 2012-02-01 00:00 3 380
.etc.
Average Rent
Beds by Suburb Jan-12 Feb-12 Mar-12 Apr-12 May-12 Jun-12 Jul-12
The Brook
1 $150 $245 $160 $285 $135 $370 $350
2 $330 $340 $380 $310 $335 $345 $355
3 $350 $380 $310 $395 $380 $350 $350
Inner City
1 $150 $245 $160 $285 $135 $370 $350
2 $330 $340 $380 $310 $335 $345 $355
3 $350 $380 $310 $395 $380 $350 $350
Any pointers or assistance would be greatly appreciated.

Answer 1

Assuming that id is the primary key of each table, then according to your query text, a property is associated with a region, by virtue of the region_id column on the properties table:
FROM properties p
LEFT
JOIN regions r
ON p.region_id=r.id
A district is associated with a region (presumably, a district is a subdivision of a region.)
LEFT
JOIN districts d
ON d.region_id=r.id
and a suburb is associated with a district (presumably, a suburb is a subdivision of a district.)
LEFT
JOIN suburbs s
ON s.district_id=d.id
The net result is that every property within a region is getting associated with EVERY district within that region, and associated with EVERY suburb within each district.
So, you are getting the rent values averaged for all properties within a region.
To get rent values per suburb, you really need the relationship between a property and its suburb.
What you really need is a suburb_id column on the properties table as a foreign key to the suburbs table.
LEFT
JOIN suburbs s
ON s.district_id=d.id
AND s.id = p.suburb_id