如何在多对多表关系中优化查询性能

Question

I have a user table and a virtual item table, each with ID values as primary keys.

I also have a favorites table where users can store their favorite items. The favorites table maps user_id to item_id. Many users can favorite many items.

table users
user_id  bigint primary key auto_increment
...

table items
item_id  bigint primary key auto_increment
...

table favorites
user_id  bigint 
item_id  bigint 

Since I cannot create indexes on the favorites table, because there are no unique or primary keys, how do i optimize search queries, such as SELECT? Is there a better way to store the data?

Answer 1

Since I cannot create indexes on the favorites table, because there are no unique or primary keys

Huh? Of course you can create indexes. In fact, two come to mind:

• favorites(user_id, item_id)
• favorites(item_id, user_id)

These would be used to answer different questions. Respective examples are: What are the favorites of user X? What users have a favorite of Y?