Question

I have three tables:

• sales, which are purchases from a store
• items, which are all the items the store carries
• sale_items, which associate sales to the items bought in that sale. So a row in sale_items has both a sale_id and an item_id.

What is the most efficient way to find all sales that include a sale_item corresponding to a given item?

My first guess was

select sale_id, item_id, count(*)
from sales as s
join sale_items as si on si.sale_id=s.id
group by item_id
order by count(*) desc

But then if a given sale had multiple sale_items with the same item, that sale is counted multiple times. That’s not what I want–I only want to count the sales themselves.

What is the best way to do this?

Answer 1

Sounds like you want to find the number of sales that included a specific item:

SELECT count(distinct(sale_id)) FROM sale_items
WHERE item_id='<your_item_id>'