Using Unique in database vs doing so programmatically

Question

I have a general database programming theory question that I hope isn’t too off base for stack overflow. One of the tables in my database holds a uniques string. I have indicated in MySQL that this field is Unique. When trying to add the same data again - on purpose - I get the proper Duplicate entry message and the data is not added. My question is, is it better to use this database feature, uniqueness, versus inserting a few statements in my code to look up the string and check if the string has been saved previously...

Thanks in advance for your help.

Jeff

Answer 1

No.

What if someone else adds it between your if check and your INSERT query?

Part of the point of database-level constraints is to be atomic – to check the constraint and perform the insert at the same time, so no-one else can interfere in the middle.

Answer 2

I would say in the case of a relational database like MySQL, it is absolutely the right decision to have the database enforce the uniqueness of a column.

Enforcing data integrity is a main feature of databases, and unique columns are an integral part of this ability. In the case of a key column, there are also performance benefits of looking up records by id when the enforcement lives in the db.

Finally, when you're running a distributed application where many users can simultaneously create records, you get the unique guarantee for free, instead of having to synchronize key creation yourself.