Very slow subquery in MySQL

Question

I have two tables A and B (names are abstracted). Both table A and table B have a column called date. Table A has about 3 million rows, with column date indexed. Table B has about 1 million rows.

I have the following query running very slowly (about 20 seconds):

SELECT * FROM A WHERE A.date > (select MAX(date) FROM B)

However, if I run the subquery separately first, it's very fast. And if I substitute the date into the main query, it's also super fast as the data in table A is indexed.

I have seen that this could be an issue if "IN" clause is used. But my query doesn't use IN. Can someone help explain why this is slow?

Thank you!

Answer 1

Move your subquery into the from clause as a derived table and join it to table A without a join condition:

SELECT * FROM A, (select MAX(date) AS mdate FROM B) AS T WHERE A.date > T.mdate

This way you can be sure that the subquery is calculated only once. If A.date is indexed as you wrote, this should speed-up your query. In the current form, I believe the subquery is executed for each record separately.

Answer 2

How many time take select MAX(date) AS mdate FROM B alone?

You can probably save a lot of second if you have an index in B .date too