Question

I’m working on a restaurant CMS app. I have a many-to-many relationship between 2 tables, menu_sections and menu_items. The relationship is maintained with a table in between called menu_relationships.

As an example let’s say the menu section called Snacks (menu_section_id = 1) contains a menu item called Pretzels (menu_item_id = 1) and the menu section called Desserts (menu_section_id = 2) contains a menu item called Ice Cream (menu_item_id = 2), but Ice Cream is also contained within another menu section called Kids Food (menu_section_id = 3). So there would be 3 rows in the menu_relationships table to map out these 3 relationships. The relationship table would look like this:

---------------------------------------
|   menu_section_id  |  menu_item_id  |
|=====================================|
|          1         |        1       |
|-------------------------------------|
|          2         |        2       |
|-------------------------------------|
|          3         |        2       |
---------------------------------------

So far so good.

I want to generate a result set that will return the names of all menu items except for menu items with a given menu_section_id. So to return the menu item names, I have a join on the menu_items table. Here’s the SQL:

SELECT menu_section_id, menu_items.menu_item_id, menu_item_name 
FROM menu_relationships
JOIN menu_items 
ON menu_items.menu_item_id = menu_relationships.menu_item_id
WHERE menu_section_id != 2 

The result set which will give me a row for each relationship that doesn’t contain a given menu_section_id. With the example data I would be getting 2 rows back from the relationship table:

-----------------------------------------------------------
|   menu_section_id   |  menu_item_id  |  menu_item_name  |
|======================================|==================|
|          1          |        1       |     Pretzels     |
|--------------------------------------|------------------|
|          3          |        2       |     Ice Cream    |
-----------------------------------------------------------

But what I want is to exclude the menu item altogether from the result set, if it has ANY relationship to the specified menu_section_id. In other words, in the case of this example , I only want to return rows for menu items that have no relationship mappings at all to a menu_section_id of 2, I only want to return the Pretzels row.

I’ve tried various things with GROUP BY and HAVING using the bit_xor() aggregate function, but so far no luck at all in getting what I want.

I probably could have taken less time to explain that but I wanted it to be a clear as I can make it. I hope it is. Can anyone help?

Answer 1

There are many ways to do this. Here is one example using WHERE ... NOT IN (...):

SELECT
    R.menu_section_id,
    I.menu_item_id,
    I.menu_item_name
FROM menu_items AS I
JOIN menu_relationships AS R
ON R.menu_item_id = I.menu_item_id
WHERE I.menu_item_id NOT IN
(
     SELECT menu_item_id  
     FROM menu_relationships
     WHERE menu_section_id = 2 
)