trig substitution is the most instructive way of doing it although the complex number way suggested by dextercioby is probably a good trick to learn.
Way of thinking:
you have something you can't do, look nothing like what u used to seeing, so you need something new, some subtle substitution...
why trig substition?
simply because you have something in the denominator [tex]1+y^2[/tex] which looks like one side of a trig identity! Observe that [tex]1-y^2[/tex] also looks like one side of a trig identity too. Now you go to the book and look at all your trig identities between [tex]\sin, \cos, \tan, \cot, \sec, \csc[/tex], which one do you think it could be useful? Remember we have [tex]y^2[/tex], so...?
now once you have picked the right one (and change of variable, that's why it is called trig substitution), you do the integrals (typically in terms of just a combination of [tex]\sin, \cos, \tan, \cot, \sec, \csc[/tex]. But if you pick the right identity to start with, the identity itself will help you simplifies this expression and it should be quite easily done given that you know how to do these integrals with trig functions. Final step is to convert everything back to the variable [tex]y[/tex] and that's where you get your inverse tan...