Keep all the observations that have maximum value of a variable in each group

[Cameron Flame]

Dear Statalist,

Here is an example of my data, two of its variables are CaseID and Surveyyear:

CaseID	Surveyyear
1	2004
1	2004
1	2004
2	1999
2	1999
2	1999
2	2000
2	2000
2	2000
3	2002
3	2002
3	2009
3	2010
3	2010

For each group of caseID that has more than 1 value of Surveyyear, I want to keep observations that have the maximum value of Surveyyear. If there is only 1 value of Surveyyear in a group, I'll keep all of the observations in that group. It means, I want to generate this following result:

CaseID	Surveyyear
1	2004
1	2004
1	2004
2	2000
2	2000
2	2000
3	2010
3	2010

Could someone help me? I tried using 'collapse' and, 'keep if _n == _N', but only one observation in each group is kept, not all the observations that have the same value.
Thank you so much in advance.
Best regard,
Cameron.

[Felix Bittmann]

bysort CaseID: egen maxyear = max(Surveyyear)
keep if Surveyyear == maxyear

[Nick Cox]

bysort CaseId (SurveyYear): keep if Surveyyear==Surveyyear[_N]

[Cameron Flame]

Both ways work well. Thank you guys a lot. :D