Mysql8.0 learning record 20 - trigger

What is trigger ？
Triggers are associated with specific tables , And the named database object triggered according to the specific event of the table .
Its creation syntax is as follows ：

CREATE
    [DEFINER = user]
    TRIGGER [IF NOT EXISTS] trigger_name
    trigger_time trigger_event
    ON tbl_name FOR EACH ROW
    [trigger_order]
    trigger_body

trigger_time: { BEFORE | AFTER }

trigger_event: { INSERT | UPDATE | DELETE }

trigger_order: { FOLLOWS | PRECEDES } other_trigger_name

The same schema The name of the trigger in must be unique .

From the definition syntax of trigger, we can see , The corresponding three types of events triggered ：INSERT、UPDATE、DELETE. This is not just for insert/update/delete sentence ：

• INSERT： When a new row is added , image INSERT\LOAD DATA\REPLACE Statements can trigger
• UPDATE： When a row of data is modified , Generally speaking, it means update sentence
• DELETE： When a row of data is deleted , It's usually delete or replace sentence . Be careful DROP TABLE or TRUNCATE TABLE It doesn't trigger delete event .

Each of the above events has two trigger opportunities :BEFORE and AFTER . In triggers , Can pass OLD and NEW Refers to the value before modification and the new value ,insert Events cannot be used OLD,delete Events cannot be used NEW.OLD The value of is read-only and cannot be modified , But it can go through set modify NEW Value . It should be noted that , Only BEFORE To modify the insert or update New value of .AFTER no way , Because the data modification has been completed .

There's another thing to pay attention to here ,NEW and OLD Cannot reference generated column .

If multiple triggers with the same event and the same timing are defined on the same table , So by default , Create first, execute first . If there is order dependency , So you can go through FOLLOWS and PRECEDES Indicates that it is executed after or before other triggers .

Example 1 of trigger ： Count the total number of large tables
Encountered a particularly large table , Every time the total number of statistics will be particularly stressful . In this case, triggers can be used to count ：

create table t_count(
	id int primary key auto_increment,
	name varchar (10),
	num int
);

create table t_count_sum(
	total long
);
insert into t_count_sum values(0);

--  Every new row of data , Counting plus one 
create trigger t_record_sum_plus after insert on t_count
for each row
update t_count_sum set total = total+1;
--  Every row of data deleted , Count minus one 
create trigger t_record_sum_subtract after delete on t_count
for each row
update t_count_sum set total = total-1;

Test it ：

insert into t_count(name,num) values('a',1),('b',2),('c',2);
delete from t_count where name = 'b';

select * from t_count_sum;

total|
-----+
2    |

Example 2 of trigger ： Count the total number of new data in a reply and the average value of a field
Still used t_count This table , Refer to the example on the official website ：

delimiter //
create trigger t_num_avg after insert on t_count
for each row 
begin 
	set @sum = @sum + NEW.num;
	set @total = @total+1;
	set @avg = @sum/@total;
end
//
delimiter ;

test ：

set @sum = 0;
set @total = 0;
insert into t_count(name,num) values('a',1),('b',2),('c',3),('d',4);
select  @avg;

insert into t_count(name,num) values('e',4);
select  @avg;

@avg       |
-----------+
2.500000000|

@avg       |
-----------+
2.800000000|

Conditions related to the failure of the trigger

• If one BEFORE Trigger failed , Then the operation of the response line will not be executed
• BEFORE Triggers are used when an operation attempts to add or modify data , It will be carried out , It doesn't matter whether the corresponding operation is successful
• AFTER Triggers are only relevant BEFORE And response line operations are successful before execution
• BEFORE and AFTER Failure of trigger itself , It will cause the statement that triggers the trigger to execute to fail

Other matters needing attention

• Triggers cannot use statements that explicitly or implicitly start or end a transaction , for example START TRANSACTION, COMMIT, or ROLLBACK.
• Stored procedures can be used in triggers to reuse the same code , But the corresponding stored procedure should not return data or use dynamic SQL

Function of trigger

• Can be used to implement InnoDB The statistics in the following table
• It can be used to verify and correct data
• It can be used to track users' operations on the database
• You can perform cascading updates of tables
• It can be used to replicate table data
• Data values can be calculated automatically , If the value of the data meets certain requirements , Then specific processing is carried out .