Drop (Keep) observations

[Ke Gong]

Dear experts,

I'm wondering how to drop (or keep) my observations. Below is my dataset. What I wanted to do is this:

If in a particular year there are two observations with adjustment shows "0" and "1", keep the observation with adjustment "1" while dropping the observation with adjustment "0". For example, in Year 2008, drop the first observation but keep the second.

If in a particular year, there is only one observation with adjustment shows "0", I'll just keep this observation. For example, there is only one observation with adjustment "0" in the year 2014, so just keep it.

* Example generated by -dataex-. To install: ssc install dataex
clear
firm        year  adjustment  revenue
300269 "2008-12-31" 0  184107248
300269 "2008-12-31" 1  184107248
300269 "2009-12-31" 0  254609552
300269 "2009-12-31" 1  254609552
300269 "2010-12-31" 0  349930400
300269 "2010-12-31" 1  349930400
300269 "2011-12-31" 0  512150592
300269 "2011-12-31" 1  512150592
300269 "2012-12-31" 0  534164544
300269 "2012-12-31" 1  534164544
300269 "2013-12-31" 0  585608128
300269 "2013-12-31" 1  585608128
300269 "2014-12-31" 0  970097984
300269 "2015-12-31" 0 1522596480
300269 "2016-12-31" 0 2803457792
300269 "2017-12-31" 1 3952366336
300269 "2018-12-31" 0 4053373440
300269 "2018-12-31" 1 4053373440
end

Thank you very much for helping!

[West Addison]

Try this:

assert inlist(adjustment, 0, 1)                    // verify that adjustment is always 0 or 1
bysort firm year (adjustment):  assert  _N <= 2    // verify that there are at most 2 observations per firm year
by     firm year (adjustment):  keep if _n == _N

[Ke Gong]

West Addison Your code works pretty well. Thank you very much!!