Identify dates in close proximity by group in longitudinal data

[Melk]

I have longitudinal data similar to below:

ID   GROUP   DATE          FLAG

1     A              1/1/2020       1

1     A              1/2/2020       1

1     B              1/3/2020

2     A               1/1/2020       

2     A                1/8/2020

2     B               1/4/2020       1

2     B                1/1/2020       1

I want to flag and and further remove observations for the same ID and GROUP where the dates are within 6 days of each other (plus or minus). How may I do this?

[novinosrin]

HI @Melk  Do you mean the following by any chance?

data have;
input ID   GROUP $   DATE :mmddyy10.;*          FLAG;
format date mmddyy10.;
cards;
1     A              1/1/2020       1
1     A              1/2/2020       1
1     B              1/3/2020
2     A               1/1/2020       
2     A                1/8/2020
2     B               1/4/2020       1
2     B                1/1/2020       1
;

data want;
 merge have have(firstobs=2 rename=(date=_date group=_group));
 if lag(id)=id and lag(group)=group and abs(dif(date))<=6 then Flag=1;
 else if group=_group and 0<abs(_date-date)<=6 then Flag=1;
 drop _:;
run;

proc print noobs;run;

ID	GROUP	DATE	Flag
1	A	01/01/2020	1
1	A	01/02/2020	1
2	B	01/03/2020	.
2	A	01/01/2020	.
2	A	01/08/2020	.
2	B	01/04/2020	1
2	B	01/01/2020	1