为什么mysql（不在）比没有子查询更快？

Question

I can write a query with two differents types:

SELECT * 
FROM `pictures` 
WHERE `field_id` NOT IN (SELECT `field_id` 
                         FROM `table` 
                         WHERE `confirm` = 0)  

and

SELECT * 
FROM `pictures` 
WHERE `field_id` NOT IN (12,56,435,44,25,52,876,99)  

But the second example are very faster!
What is the reason for being faster?

Answer 1

Here's a good performance rule of thumb: doing less work takes less time.

In the simpler example, the MySQL engine executes exactly 1 query:

-- Execute the main query (takes time M)
SELECT * 
FROM `pictures` 
WHERE `field_id` NOT IN (12,56,435,44,25,52,876,99)  

In the more complex example, the MySQL engine executes 2 queries:

-- Execute the subquery (takes time S)
SELECT `field_id` FROM `table` WHERE `confirm` = 0;
-- => (12,56,435,44,25,52,876,99)

-- Plug the subquery results into the main query (takes time M)
SELECT * 
FROM `pictures` 
WHERE `field_id` NOT IN (12,56,435,44,25,52,876,99);

The simpler example takes about time M in total, whereas the more complex example takes about time S + M , plus maybe a bit more overhead to plug everything, and S + M is greater than M .