Combining three tables when two have the same foreign key - mysql

Question

Hello I am making some accounting software that allows you to store transactions across many months (stored in table transactions) and categorize those transactions based on a list of available categories (stored in table categories). Finally each month the user can create a budget where they build a list (from a subset of categories) and assign a goal to each entry. These lists are stored in table budget.
I need a query that returns a spending and budget summary for a specified month. This summary would be a table of categories.name, budget.goal, and sum(transactions.amount).
Sometimes a user would have a budget item for the specified month but hasn't made any transactions with that category yet (Out to Eat example). Sometimes a user would have an unexpected expense they didn't budget for (Auto Repair Example) and there will be some categories (like vacation) where the user didn't budget that item and there were no expenses of that category.
SELECT categories.id, categories.name, SUM(transactions.amount)
FROM categories
LEFT JOIN transactions ON categories.id=transactions.category_id
WHERE transactions.date LIKE '2019-08-%' GROUP BY categories.id;
gets me half of what I want and
SELECT categories.id, categories.name, budgets.goal
FROM categories
LEFT JOIN budgets ON categories.id=budgets.category_id
WHERE budgets.date LIKE '2019-08-%' GROUP BY categories.id;
gets me the other half of what I want. Is there a single query that can return results like what I have pictured above? I would be even more thrilled if we could exclude results where both goal and sum are NULL.

Answer 1

If you know that there can be at most one budget per category and month, then you can (LEFT) JOIN both (child) tables to categories:
SELECT
c.id,
categories.name,
MAX(b.goal) as budget_goal,
SUM(t.amount) as total_txn_amount
FROM categories c
LEFT JOIN budgets b
ON c.id = b.category_id
AND b.date LIKE '2019-08-%'
LEFT JOIN transactions t
ON c.id = t.category_id
GROUP BY categories.id
HAVING COALESCE(budget_goal, total_txn_amount) IS NOT NULL;
Note that while we know that there can be only one budget per group, the engine doesn't, and might claim that b.goal must be either in the GROUP BY clause or used in an aggregate function. So we use MAX(b.goal) to avoid that error.
To improve the performance of the first JOIN I would change
AND b.date LIKE '2019-08-%'
to
AND b.date >= '2019-08-01'
AND b.date < '2019-08-01' + INTERVAL 1 MONTH
and create a composite index on (category_id, date).
Also to force the uniqueness of category and month combination in the budgets table, I would create a virtual column like
year_month VARCHAR(7) as (LEFT(date, 7))
and a UNIQUE KEY on (category_id, year_month)
Then you can use
LEFT JOIN budgets b
ON c.id = b.category_id
AND b.date = '2019-08'

Answer 2

You can try below query to find the proper result :
SELECT c.name,b.goal,sum(ct.amount) FROM era.categories c left join budget b on b.cat_id=c.id left join transactions ct on ct.cat_id=c.id WHERE b.date LIKE '2019-08-%' group by ct.cat_id;
Thanks.