Re: on duplicate key update for triggered table

[Rick James]

OK, maybe that is a good use for TRIGGERs.

If TEST3 has the same fields as TEST2, but with the addition of a TIMESTAMP FIELD at the beginning, you could also record the time at which the dup came in. They body of the trigger would be something like:

BEGIN
INSERT INTO TEST3 (a,b,c) SELECT a,b,c FROM NEW;
END;

Or... TEST3 could reference TEST2:
BEGIN
INSERT INTO TEST3 (id) SELECT id FROM NEW;
END;
This assumes that just the id (the PRIMARY KEY or some UNIQUE key -- one column or a set) is sufficient to tell you what you need. That is, I am assuming there are no other fields you want to capture. Again, TEST3 could have a TIMESTAMP field that will default to the current time.

Or... Your client code could check RowsAffected (or whatever is the right thing) to see when an UPDATE/INSERT fails to do anything (implying a dup).