Dear Statalists,

my data set looks like this:

Meeting Date Year Company Attending Person
25may2011 2011 A Bob
25may2011 2011 A John
25may2011 2011 A Peter
25may2011 2011 A Seth
25may2011 2011 A Marina
25may2011 2011 A Rachel
14nov2011 2011 A Bob
14nov2011 2011 A John
14nov2011 2011 A Peter
14nov2011 2011 A Seth
14nov2011 2011 A Marina
14nov2011 2011 A Rachel

Of course there are many companies (about 2.000). Usually there is one meeting per company per year. Some companies had more than one meeting in a year. What I would like to do, is to delete all meetings that took place earlier in a year, if in that year more than one meeting took place. For my example above, that means deleting all observations with meeting date = 25may2011, but only for company A.
Please help me out! Thanks!