views:

145

answers:

3

I have the following two tables that record expenditure and provide expenditure category information:

Table transactions:

+-------+--------+--------+
| month | cat_id | amount |
+-------+--------+--------+
|     1 |      2 |      3 |
|     1 |      2 |      8 |
|     2 |      1 |      7 |
|     2 |      1 |      5 |
+-------+--------+--------+

Table categories:

+--------+-------------+
| cat_id | cat_desc    |
+--------+-------------+
|      1 | Stock       |
|      2 | Consumables |
+--------+-------------+

What I would like is to construct a query that displays a sum of the amounts for each category, for each month, even if there is no expenditure in that category for that month like this:

+-------+-------------+--------+
| month | cat_desc    | amount |
+-------+-------------+--------+
|     1 | Stock       |      0 |
|     1 | Consumables |     11 |
|     2 | Stock       |     12 |
|     2 | Consumables |      0 |
+-------+-------------+--------+

I suspect an outer join would need to be used but I haven't found a statement to do it yet.

Thank you for any help.

A: 
SELECT c.cat_id, c.cat_desc,t.month, SUM(t.amount) 
FROM categories c 
LEFT JOIN transactions t ON (t.cat_id = c.cat_id)
GROUP BY c.cat_id,t.month
a1ex07
Thanks for the reply, but it didn't work. Using your SQL, this is what I get:1 | Stock | 2 | 12 |2 | Consumables | 1 | 11 |Sorry about the formatting
Damo
+2  A: 

This one should provide you with the correct result. The inner select prepares a list of all months combined with all categories, and the LEFT JOIN handles the rest.

SELECT t.month, t.cat_desc, COALESCE(SUM(t2.amount), 0) amount
FROM (
  SELECT DISTINCT t.month, c.cat_id, c.cat_desc
  FROM categories c
  CROSS JOIN transactions t
) t
LEFT JOIN transactions t2 ON ( t2.month = t.month AND t2.cat_id = t.cat_id )
GROUP BY t.month, t.cat_desc

Performance might be better with the following (using DISTINCT only where necessary), but you will have to try:

SELECT t.month, t.cat_desc, COALESCE(SUM(t2.amount), 0) amount FROM (
  SELECT t.month, c.cat_id, c.cat_desc
  FROM
  (SELECT DISTINCT month FROM transactions) t
  CROSS JOIN categories c
) t
LEFT JOIN transactions t2 ON ( t2.month = t.month AND t2.cat_id = t.cat_id )
GROUP BY t.month, t.cat_desc
Peter Lang
Brilliant - thank you. It works perfectly.
Damo
A: 
SELECT c.cat_id, c.cat_desc,t.month, SUM(t.amount)  
FROM categories c  
LEFT JOIN transactions t ON (t.cat_id = c.cat_id) 
GROUP BY t.month,c.cat_id 
Order By t.month
Jonathan