How to check if an existing row is identical before inserting a new one?

Question

MySQL

Good afternoon, please tell me if it is possible to put additional logic in INSERT, or do you need separately logic first, and then the action

There is a table

id ; login ; host ; event_id ; date
1; strelkov.av ; home_pc ; 1 ; 11:00:00
2; strelkov.av ; home_pc ; 2 ; 12:00:00
3; strelkov.av ; home_pc ; 3 ; 13:00:00
4; strelkov.av ; home_pc ; 1 ; 14:00:00

Now the command will go

insert into table (login, host, event_id, date) values ('strelkov.av', 'home_pc', '1', now());

And the bottom line is that, in theory, in the context of the login - strelkov.av and the host - home_pc , the last entry in the table already has an event_id, then in this case you need to skip insert and not do it. when inserting a new value, first check for the same login and host data, if the last record has the same event_id as the new one being inserted, then do not insert
if the last record has a different event_id, then you can safely insert a new

record Substitute logic in the request, or it needs to be done separately by a procedure, or on the client side, first check

Answer 1

INSERT INTO my_table (login, host, event_id) 
SELECT 'strelkov.av', 'home_pc', 1
FROM my_table
WHERE login='strelkov.av' 
AND host='home_pc'
AND event_id=1
HAVING COUNT(*)=0;

Please note that you need to specify login, host and event_id values ​​in 2 places. It works like this - if the SELECT ... HAVING COUNT(*)=0 construct returns true (that is, SELECT did not find a record with the same login, host, and event ID), then INSERT will be performed. If SELECT returns false (COUNT is not equal to 0, then such a row already exists), then INSERT will be ignored entirely. I understand that this does not immediately get into it, but SQL is an interesting thing :)
PS: It is assumed that your id is AUTO_INCREMENT, while date has DEFAULT NOW () and you don’t need to pass them at all.