I have a schema in the effect of the following, simplified to illustrate my point:

1. users table
2. bank_deposits table – belongs to user
3. orders table – belongs to user
4. transactions table – which contains debit/credit transactions

Where transactions is related to either bank_deposits OR orders but not both at the same time — and so related to users through them.

I want a simple way to get ALL transactions belonging to the user (to calculate balance) but adding a user_id column directly in transactions would be redundant.

In reality, there are more tables between bank_deposits/orders and transactions so I really want to avoid complicated the queries.

Should I add a redundant user_id column in transactions for efficiency sake?