MySQL SUM distinct rows based on foreign key without subquery

Question

I have the following MySQL tables, where B_ID is a foreign key pointing to the ID column of Table_B:

----------    -----------
|Table_A |    |Table_B  |
----------    -----------
|B_ID int|    |ID int   |
----------    |VALUE int|
              -----------

I want to compute the sum of the VALUE column in Table_B considering only IDs that appear as foreign keys in Table_A. Also, I want to add each row in Table_B at most once.

So, if the tables were filled with the following data:

------    ------------
|B_ID|    |ID | VALUE|
------    ------------
|1   |    |1  | 50   |
|2   |    |2  | 100  |
|2   |    |3  | 200  |
------    ------------

the sum would be 150 (computed as 50+100).

The query I am using now is:

SELECT SUM(b.VALUE)
FROM Table_B b
WHERE b.id IN (SELECT DISTINCT B_ID FROM Table_A)

However, I would like to avoid having a subquery if at all possible. Any ideas on how to do this without subqueries?

Note: This query is itself a subquery in a larger query, so it can only select one column.

Answer 1

This does not eliminate subqueries, but would be more efficient if you have few records in a and lots in b :

SELECT  SUM(value)
FROM    (
        SELECT  DISTINCT b_id
        FROM    a
        ) a
JOIN    b
ON      b.id = a.b_id

Create an index on a.b_id for this to work fast.

If there are few records in b and lots of them in a , your original query will be faster. The suggestion of the index still applies.