If the limit exists (and I haven't gone so far as to prove that) then it must exist for x, y --> (0, 2) from any way of approaching the point (0, 2). What I would do is to take, say, y = x + 2. Then after you sub y = x + 2 into the limit, you evaluate it for x --> 0.

-Dan