Mysql nested left joins with counts - mysql

Question

I have a set of MySQL three tables in a "has many" relationship: deals, orders, and coupons.
Deals
|----|--------------|
| id | title |
|----|--------------|
| 1 | Some deal |
| 2 | Another deal |
|----|--------------|
Orders
|----|---------|-----------|
| id | deal_id | state |
|----|---------|-----------|
| 1 | 1 | purchased |
| 2 | 1 | purchased |
| 3 | 1 | expired |
| 4 | 2 | purchased |
|----|---------|-----------|
Coupons
|----|----------|
| id | order_id |
|----|----------|
| 1 | 1 |
| 2 | 1 |
| 3 | 1 |
| 4 | 2 |
| 5 | 2 |
| 6 | 4 |
|----|----------|
So, deals have many orders, which have many coupons.
What I'd like to do is select on the deals table while counting the number of purchased orders and coupons.
I know how to get a count on paid orders already:
SELECT deals.*, count(orders.id) AS orders_purchased_count FROM deals
LEFT JOIN orders ON deals.id=orders.deal_id AND orders.state='purchased'
WHERE deal_id < 3
GROUP BY deals.id
Deals
|----|--------------|------------------------|
| id | title | orders_purchased_count |
|----|--------------|------------------------|
| 1 | Some deal | 2 |
| 2 | Another deal | 1 |
|----|--------------|------------------------|
Similarly, I can get a count of coupons for orders:
SELECT orders.*, count(coupons.id) AS coupons_count FROM orders
LEFT JOIN coupons ON orders.id=couoons.orders_id
WHERE orders.state='purchased'
GROUP BY orders.id
Orders
|----|-----------|---------------|
| id | state | coupons_count |
|----|-----------|---------------|
| 1 | purchased | 3 |
| 2 | purchased | 2 |
| 4 | purchased | 1 |
|----|-----------|---------------|
My question is: How do I combine these so that I can add coupons_count next to orders_purchased_count?
Deals
|----|--------------|------------------------|---------------|
| id | title | orders_purchased_count | coupons_count |
|----|--------------|------------------------|---------------|
| 1 | Some deal | 2 | 5 |
| 2 | Another deal | 1 | 1 |
|----|--------------|------------------------|---------------|
The tricky thing, in my case, will be to run the WHERE deal_id < 3 filter when selecting from deals before I join on orders and to run the WHERE orders.state='purchased' filter when selecting from orders before I join on coupons. It's a large dataset and I don't want to load all my orders and coupons into memory for the purpose of joining.
At a loss for how to do this.

Answer 1

Does this work? I couldn't understand your concerns which part was going to be tricky.
SELECT deals.*, COUNT(DISTINCT(orders.id)) AS orders_purchased_count, COUNT(coupons.id) AS coupons_count
FROM deals
LEFT JOIN orders
ON deals.id=orders.deal_id
AND orders.state='purchased'
LEFT JOIN coupons
ON orders.id=coupons.orders_id
WHERE deal_id < 3
GROUP BY deals.id;

Answer 2

Try this one with a co related subquery and join
SELECT deals.*, count(orders.id) AS orders_purchased_count
(
SELECT count(id) FROM coupons WHERE orders_id = orders.id
) AS coupons_count
FROM deals
LEFT JOIN orders ON deals.id=orders.deal_id AND orders.state='purchased'
WHERE deal_id < 3
GROUP BY deals.id