Inserting data with a unique good is still high efficiency with COUNT (*) query before inserting?

Is it better to insert data with a unique or count (*) query before inserting?
A table does not allow inserting multiple title + URLs with the same data.

Mode A: With a UNIQUE constraint, the code simply executes the statement insert ignore into ....

Way B:select Count (*) from mytable where title=? and url=?
The result is zero to execute insert INTO ....



Which is the most efficient way of doing it?
------Solution--------------------
How is the first way efficiency is high ...

------Solution--------------------
Agreed upstairs, but not tested.

