Question

CREATE TRIGGER `after_customer_insert`
AFTER INSERT ON `customer`
FOR EACH ROW BEGIN
    UPDATE `user`
    SET
        `customers_count` = `customers_count` + 1
    WHERE `id` = NEW.`user_id`;
END$$

Instead of calling UPDATE user for each new row in customer, would be possible to have a trigger as a “whole”? I mean something like

CREATE TRIGGER `after_customer_insert`
AFTER INSERT ON `customer`
BEGIN
    UPDATE `user`
    SET
        `customers_count` = (
            SELECT COUNT(`id`)
            FROM `customer`
            WHERE `user_id` = `id`
        )
END$$

Answer 1

Triggers are mainly intended to be used to execute actions related to a change on a single row, usually changing another single row, as Jake points out. Of course you can implement logic that (redundantly) carries out the same action on largely the same result set for each row updated, but the question is, why would you not issue a single UPDATE statement subsequent to your original UPDATE? You may have a good reason for this, hence Jake’s question about the overall situation.

e.g. (syntax not tested, but the general idea holds)

UPDATE u set u.customer_count = x.newcount
from
user u inner join 
(
select user, count(*) as newcount
from user group by customer_id
) x