It's a handy option when you get half way through your database creation and something fails (ran out of space, syntax issue, etc.) and you need to re-run it... the previous datafiles have been created, so you just want to "reuse" them.

In some environments datafiles take a long time to create. For example 8.0.x unicode, 500 mb datafile, on HP takes 20 minutes. Reusing an existing datafile is very handy in this scenario.