keeping the first observation and the first outcome occurrence

[haytham sheerah]

I have a data set that I would like your help in it
The data is for general population how have been following up in a hospital ( 100,680 obs)

Step 1 : There is some individuals who came only once and those I would like to drop them ( I think this is the code )
bysort ID: drop if _N==1

Step 2 : I want to keep the 1st observation when a disease had occur

for example

ID	exam day	DM
1	20200601	0
1	20201201	0
1	20210201	1
2	20190101	0
2	20190202	0
2	20190601	1
2	20191001	0
2	20200202	1
3	20180606	1

I would like to delete the 2nd row for ID 1, rows 2 4 5 for ID 2, and ID 3

ID	exam day	DM
1	20200601	0
1	20210201	1
2	20190101	0
2	20190601	1

waiting for your positive reply

[Nick Cox]

The lack of replies here may arise from your not using dataex which as explained at https://www.statalist.org/forums/help#stata is especially important for date variables.

Your date variable could be (a) a string variable (b) a long integer (c) a Stata daily date variable formatted as is. We just can't tell from your example. exam day is not a legal variable name

This may help. See also https://www.stata.com/support/faqs/d...t-occurrences/

clear 
input ID long exam_day    DM
1    20200601    0
1    20201201    0
1    20210201    1
2    20190101    0
2    20190202    0
2    20190601    1
2    20191001    0
2    20200202    1
3    20180606    1
end 

gen date = daily(string(exam_day, "%8.0f"), "YMD")
format date %td 

bysort ID (date) : drop if _N == 1 
bysort ID (date) : keep if sum(DM) == 1 
bysort ID (date) : keep if _n == 1 

list, sepby(ID)