How can mysql update be executed after the request?

As many have said, you do not need it.

The reason for this question is the lack of understanding of the elementary bases of databases.

The value should not be checked by the script that writes the data, but by the one that reads.

That is, in this case, we do not write a specific value to the database at the right time, but write the desired time for changing the value.

And all reading scripts will read this time and compare with the current one. If more, then the value has changed. Very simple, and does not require any crutches from other answers. This is how databases all over the world work.

1. The WHERE clause in the SQL query only specifies a condition that will satisfy the record if it is met. It does not control the execution time of the request in any way.
2. Where did you get this syntax from?
3. Writing a query like this is an extremely bad practice. Learn how mysqli_stms_bind_param works .
4. In order to postpone the execution of any team there are a lot of options, but all of them are crutches to some extent.

For example:

• You can create a JOB in mysql, which will periodically read the queue table (in which you will enter update tasks) and perform those that are older than a minute
• You can stupidly slow down the execution of the script operator sleep(60)
• It is possible by curl (with session break) to call a script that will perform the update in a minute
• You can save tasks for update to a file (or database table) and start processing by cron (almost the same as the first option)
• You can ponder with events
• You can look at the FPM and its fastcgi_finish_request function
• Etc. etc.

1. Option 1 - use the mysql SLEEP () function, DO SLEEP(60);

   An example mysql query:
   DO SLEEP(60);
UPDATE ...
   you can use SELECT SLEEP(60) , but in this case you need to consider that the output will always be 0 or 1.

2. Option 2 - use the php sleep () function.

   Example:
   SLEEP(60);
$this->db->query("...");

Both options delay execution for N seconds.

These functions should not be used if you execute the next request only after receiving the results of the previous one, because, for example, a consecutive call of 2 requests will result in an idle time of 2 * N seconds.