Best Method Of Checking For Duplicate Entries In SQL Server

Here is my situation.  I have a table in my application that pairs users with cars they like.  We'll call this table Favorites.  A user can browse the site and they can designate as many cars they want as favorites.  For example, a user can go to the Honda Accord page and add that as a favorite car and then go to the Toyota Camry page and add that as a favorite car.  However, if he/she goes to that Honda Accord page and tries to click the "Add to Favorites" button again, at the present state of my application, it will just add another entry into the Favorites table with a duplicate pairing.  So, if I were to datalist the table to generate a listing of all favorites belonging to a certain user, he/she may potentially be returned with superfluous duplicate entries.  Not to mention, taking up valuable database space and not looking very professional.

In my Favorites table, the 3 fields are.....

favoriteId (set as primary key)
userId
carId

I've been thinking about this for awhile and I've come up with 2 solutions.  I'm a newbie to ASP.NET/programming so I don't have enough insight to make a decision or to even think up of other alternatives.

1) Check proactively by doing a.....
SELECT favoriteID FROM Favorites WHERE userId = x and carId = y  (where x and y are variables)
If I get a null return, it means I can go ahead and let the user add the car as a favorite in the database.  If I get a valid value, then it means there already exists the same pairing, so I exit out without updating the table.

2) Check reactively by forcing an exception whenever a user tries to enter a duplicate pairing.  I'm not sure how to do this, but perhaps, instead of making "favoriteId" a primary key, perhaps, I can make a primary key pairing of "userId" and "carId".  And by trying to do an insert with a primary key that already exists, we know it won't work since primary keys by definition are unique.

Now, I expect some concurrent users on my site, so I must take into consideration pros and cons of each and determine which is more efficient.  Checking proactively will force a check even if the table does not contain a duplicate pairing of user and car.  However, having a duplicate primary key may be more expensive from a database point of view and may slow down lookups, etc.  Or maybe neither has significant benefits, in which case, I rather go with proactive, since I've already coded it and it works fine.  Or maybe there is a third alternative, which I did not think.  Which method do programmers usually take and which is a better practice?

TIA for your help.