Join distant SQL tables without pulling data in between - sql

Question

I have two mySQL tables that are distantly related: Box and Schedule.
The tables are only related through many other table joins.
A box has a subscription ID, a subscription has a boxContent ID and a boxContent has a Schedule ID.
Many boxes have one subscription
One Subscription has many boxContents
Many boxContents have one Schedule
I need to get all of the Schedule details attached to boxes and group them by the delivery date of the boxes.
The problem I have, is that I cannot figure out how to group them, or join them, so that the subscription and boxcontents data doesnt fill the results up with invalid results.
For example:
SELECT
Box.expected_delivery_date, count(Box.id) num_boxes
FROM
Box
WHERE
Box.state = 3 AND Box.status = 2
GROUP BY Box.expected_delivery_date;
returns the correct amount of boxes that I need to look for:
2010-10-01 - 2
2010-10-07 - 4
However, when I include the tables I need to link the two together, I get all of their data confusing the results
SELECT
Box.expected_delivery_date, count(Box.id) num_boxes
FROM
Box
JOIN
Subscription ON Box.subscription_id = Subscription.id
JOIN
BoxContent ON Subscription.id = BoxContent.subscription_id
JOIN
Schedule ON Schedule.id = BoxContent.schedule_id
WHERE
Box.state = 3 AND Box.status = 2
GROUP BY Box.expected_delivery_date;
Returns:
2010-10-01 - 10
2010-10-07 - 17
Im sure there is a simple way to do this, Im just not great at SQL.

Answer 1

Use DISTINCT to count the distinct Box.id in your query -
SELECT
Box.expected_delivery_date, count(DISTINCT Box.id) num_boxes
FROM
Box
JOIN
Subscription ON Box.subscription_id = Subscription.id
JOIN
BoxContent ON Subscription.id = BoxContent.subscription_id
JOIN
Schedule ON Schedule.id = BoxContent.schedule_id
WHERE
Box.state = 3 AND Box.status = 2
GROUP BY Box.expected_delivery_date;
This should return
2010-10-01 - 2
2010-10-07 - 4
Similarly, when you JOIN box with subscription, content, schedule tables you will get many duplicates. You need to analyze the data and see how you need to GROUP BY.
Use this query to see the actual data used by the query before grouping and decide on which columns to group by. Mostly, it will be the columns where you see duplicate data in multiple rows.
SELECT
Box.expected_delivery_date, Box.id BoxID, Schedule.id SchID
FROM
Box
JOIN
Subscription ON Box.subscription_id = Subscription.id
JOIN
BoxContent ON Subscription.id = BoxContent.subscription_id
JOIN
Schedule ON Schedule.id = BoxContent.schedule_id
WHERE
Box.state = 3 AND Box.status = 2
You may even try SELECT Box.*, Schedule.* in above query to come up with a final grouping.
If you need any more specific answer, you will have to provide the dummy data for all those table and the result you are looking for.

Answer 2

I would agree with the comments and try using DISTINCT. This should filter out the duplicate rows you would be getting by joining your "middle" tables.